Ecological Diagnostics for Marine Mammals: Appraisal

of molecular-based methods for dietary and age

estimation

Glenn John Dunshea

BSc (Hons) GCRC

Submitted in fulfillment of the requirements for the degree of

Doctor of Philosophy

Institute of Marine and Antarctic Studies

University of Tasmania

April 2012

“The ballast of factual information, so far from being just about to sink us, is growing less

daily. The factual burden of a science varies inversely with its degree of maturity”

Sir Peter Medawar, OM. “Two conceptions of science”,

Henry Tizard Memorial Lecture, Encounter 143, August 1965

Declarations

Originality;

This thesis contains no material which has been accepted for a degree or diploma by the

University or any other institution, except by way of background information and duly

acknowledged in the thesis, and to the best of the my knowledge and belief no material

previously published or written by another person except where due acknowledgement is

made in the text of the thesis, nor does the thesis contain any material that infringes

copyright.

Signed: Dated:

Glenn J Dunshea

Access;

This thesis may be made available for loan. Copying and communication of any part of this

thesis is prohibited for two years from the date this statement was signed; after that time

limited copying and communication is permitted in accordance with the Copyright Act 1968.

Signed: Dated:

Glenn J Dunshea

Ethics;

The research associated with this thesis abides by the international and Australian codes on

human and animal experimentation, the guidelines by the Australian Government's Office of

the Gene Technology Regulator and the rulings of the Safety, Ethics and Institutional

Biosafety Committees of the University.

This research was approved by the Animal Ethics Committee of the University of Tasmania

(Permit # A0008315).

Signed: Dated:

Glenn J Dunshea

Peer reviewed literature already published from parts of this thesis

Paper 1:

Dunshea, G., N. B. Barros, R. S. Wells, N. J. Gales, M. Hindell, A. and S. N. Jarman (2008)

Pseudogenes and DNA-based diet analyses; a cautionary tale from a relatively well sampled

predator-prey system. Bulletin of Entomological Research 98: 239-248.

Paper 2:

Dunshea, G. (2009). DNA-based diet analysis for any predator. PLoS One 4(4):

e5252.doi:10.1371/journal.pone.0005252.

Paper 3:

Garde E, A.K. Frie, G Dunshea , SH Hansen, KM Kovacs, C Lydersen (2010) Harp seal ageing

techniques-teeth, aspartic acid racemization, and telomere sequence analysis. Journal of

Mammalogy, 91: 1365–1374.

Paper 4:

Dunshea, G., D. Duffield, N. Gales, M. Hindell, R. S. Wells and S. N. Jarman (2011).

Telomeres as age markers in vertebrate molecular ecology. Molecular Ecology Resources, 11:

225-235.

Statement of Co-Authorship

The following people and institutions contributed to the publication of the work undertaken

as part of this thesis:

Paper 1 G. J. Dunshea (70%), N. B. Barros (5%), R. S. Wells (5%) N. J. Gales (5%)

M. A. Hindell (5%), S. N. Jarman (10%)

Paper 3 G. J. Dunshea (20%), Garde E (35%), A.K. Frie (25%), S.H. Hansen (10%),

K.M. Kovacs (5%), C Lydersen (5%)

Paper 4 G. J. Dunshea (65%), D. Duffield (5%), N. Gales (5%), M. Hindell (10%),

R. S. Wells (5%), S. N. Jarman (10%)

Details of the Authors roles:

Paper 1

G. J. Dunshea contributed to the idea of the manuscript, its formalisation and development,

performed the laboratory work, data collection, data analysis and wrote the paper

S. N. Jarman contributed to the idea of the manuscript, its formalisation, development and

manuscript refinement and presentation

N. B. Barros and R. S. Wells assisted with sample collection and manuscript refinement and

presentation

N. J. Gales and M. A. Hindell assisted with manuscript refinement and presentation

Paper 3

G. J. Dunshea, contributed to the idea of the paper, provided harp seal telomere assay data,

data analysis and interpretation, wrote the telomere part of the paper and assisted with

manuscript refinement and presentation

Garde E. contributed to the idea of the paper, its formalisation and development , provided

aspactic acid racemization data, its analysis and interpretation and wrote the bulk of the

paper

A.K. Frie and S.H. Hansen contributed to the idea of the paper, its formalisation and

development , provided teeth growth layer group analysis, sample collection and manuscript

refinement and presentation

K.M. Kovacs and C Lydersen contributed to the idea of the paper, its formalisation and

development, performed data analysis and manuscript refinement and presentation

Paper 4

G. J. Dunshea, contributed to the ideas in the paper, their formalisation and development,

performed most of the laboratory work, data collection, data analysis and wrote the paper

D. Duffield assisted with providing extracted DNA from historically archived samples

N. Gales contributed to the ideas in the paper, their formalisation, development and

manuscript refinement and presentation

M. Hindell contributed to the ideas in the paper, their formalisation, development and

manuscript refinement and presentation

R. S. Wells assisted with sample collection and manuscript refinement and presentation

S. N. Jarman contributed to the ideas in the paper, their formalisation, development and

manuscript refinement and presentation

We the undersigned agree with the above stated “proportion of work undertaken” for each

of the above published (or submitted) peer-reviewed manuscripts contributing to this thesis:

Signed: __________________ ______________________

Mark Hindell Mike Coffin

Supervisor Director

Institute of Marine and Antarctic Studies Institute of Marine and Antarctic

Studies

University of Tasmania University of Tasmania

Date:_____________________

Table of Contents Acknowledgements .......................................................................................................................... 1

Abstract ............................................................................................................................................ 2

- Introduction .................................................................................................................. 7 Chapter 1

1.1 Diet Investigations ................................................................................................................. 7

1.2 Age-based Investigations ..................................................................................................... 10

1.3 Structure of the thesis ......................................................................................................... 13

- A universal DNA-based diet technique for application to potentially any predator .. 15 Chapter 2

2.1 Introduction ......................................................................................................................... 15

2.2 Materials and Methods ........................................................................................................ 16

2.2.1 In silico development .................................................................................................... 16

2.2.2 Development of laboratory protocol ............................................................................ 19

2.2.3 Data analysis ................................................................................................................. 22

2.3 Results .................................................................................................................................. 23

2.3.1 Taxon specific presence/absence of Pac I restriction site in amplicon ......................... 23

2.3.2 ‘Universal’ Complementarity of PCR primers ............................................................... 24

2.3.3 Assessment of target 16SPLSU fragment to identify species from sequence data ...... 27

2.3.4 SSCP scoring of identical clones .................................................................................... 30

2.3.5 Removal of predator amplicons .................................................................................... 30

2.3.6 Prey detection from captive dolphin scat ..................................................................... 31

2.3.7 Prey detection from free ranging Sarasota Bay dolphin scat ....................................... 32

2.4 Discussion ............................................................................................................................. 34

- DNA-based diet analyses to describe cetacean diet; experimental and empirical Chapter 3

evaluations ..................................................................................................................................... 39

3.1 Introduction ......................................................................................................................... 39

3.2 Methods ............................................................................................................................... 40

3.2.1 Initial Tests of Efficacy of Different Protocols for Dolphin faeces DNA Extraction ....... 40

3.2.2 Contamination Control of DNA Extracts and PCR Assays ............................................. 41

3.2.3 Samples Used ................................................................................................................ 41

3.2.4 Universal Assays with SWBD samples ........................................................................... 43

3.2.5 PCR Primer Design and Testing ..................................................................................... 44

3.2.6 Real Time PCR screening for Mugil cephalus DNA in SWBD samples ........................... 45

3.2.7 Universal Assays with Sarasota Bay samples ................................................................ 45

3.2.8 Data Analysis ................................................................................................................. 45

3.3 Results .................................................................................................................................. 48

3.3.1 DNA Extraction Test ...................................................................................................... 48

3.3.2 Contamination Control of DNA Extracts and PCR Assays .............................................. 48

3.3.3 SWBD Universal Assays ................................................................................................. 49

3.3.4 SWBD Mugil cephalus DNA assays ................................................................................ 54

3.4 Sarasota Bay dolphin Universal Assays ............................................................................ 56

3.4.1 Prey Assignment by BLAST threshold and Placement in Neighbour Joining Analysis... 57

3.4.2 Prey Assignment Using the MOTU/Statistical Assignment Package method ............... 59

3.4.3 Comparison of different prey taxonomic assignment techniques ................................ 59

3.4.4 Occurrence and Inferred Relative Importance of Sarasota Bay dolphin prey .............. 64

3.4.5 Comparison Sarasota Bay dolphin diet estimated via DNA and stomach contents

data ........................................................................................................................................ 64

3.5 Discussion ............................................................................................................................. 68

3.5.1 Captive Feeding trials and SWBD prey detection results .............................................. 68

3.5.2 Sarasota bay data .......................................................................................................... 70

3.5.3 Conclusion ..................................................................................................................... 73

- Pseudogenes and DNA-based diet analyses: a cautionary tale from a relatively Chapter 4

well sampled predator-prey system .............................................................................................. 75

4.1 Introduction.......................................................................................................................... 75

4.2 Materials and Methods ........................................................................................................ 77

4.2.1 Sample collection and analysis ...................................................................................... 77

4.2.2 Sequence scoring and analysis ...................................................................................... 77

4.2.3 Confirmation of NUMT origin of spurious sequences ................................................... 78

4.3 Results .................................................................................................................................. 78

4.3.1 pNUMT frequencies, proportions and PCR characteristics ........................................... 78

4.3.2 NUMT sequence characteristics, phylogenetic analysis and substitution pattern ....... 81

4.3.3 Confirmation of NUMT origin of sequences ................................................................. 81

4.4 Discussion ............................................................................................................................. 85

– Telomeres as age markers for marine mammals ....................................................... 89 Chapter 5

5.1 Introduction.......................................................................................................................... 89

5.2 Methods ............................................................................................................................... 91

5.2.1 General Characteristics of Study Populations and Sample Sources ............................. 91

5.2.2 Bottlenose Dolphin samples ......................................................................................... 92

5.2.3 Southern Right Whales ................................................................................................. 95

5.2.4 Harp Seal Samples ......................................................................................................... 95

5.2.5 Data Analysis ................................................................................................................. 97

5.2.6 qPCR Assay Parameter Information .............................................................................. 98

5.3 Results .................................................................................................................................. 99

5.3.1 Bottlenose Dolphins ...................................................................................................... 99

5.3.2 Southern Right Whale qPCR assays ............................................................................ 105

5.3.3 Harp Seals.................................................................................................................... 105

5.4 Discussion ........................................................................................................................... 108

5.4.1 Bottlenose Dolphins and Southern Right Whales ....................................................... 108

5.4.2 Harp Seals.................................................................................................................... 110

5.4.3 Methodological Considerations .................................................................................. 112

– General Discussion ................................................................................................... 113 Chapter 6

6.1 DNA-based methods to study cetacean diet ..................................................................... 113

6.2 Live marine mammal age estimation and telomeres ........................................................ 120

6.2.1 Telomeres for age estimation: Reappraisal of the literature ..................................... 120

6.3 Conclusions ........................................................................................................................ 128

References ................................................................................................................................... 129

List of Figures

Figure 2.1 Proportion of species in alignments provided at

http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0005252#s5

(number of species in each alignment provided in parentheses – x-axis) that have priming

regions directly complimentary at the first 8- 14 nucleotide positions of the 3’ ends to the

primers presented in the main manuscript (grey bars). Hatched bars represent the same

analysis when the additional primers (5’-3’): AAGACCCCGTTGAGCTT, AAGACCCTGTCGAGCTT,

AAGACCCTATCGAGCTT, AAGACCCTTTGGAGCTT, AAGACCCTATAAAACTT,

AAGACCCTGTGGAACTT, AAGACCCTATCGAACTT, AAGACCCTATAAATCTT and

AAGACCCTATAGATCTT are included. Exact proportion is indicated above bars. ................... 25

Figure 2.2 Test of universal primers designed for the small fragment of 16s mtDNA targeted

in this study. Lanes: M : 1kb ladder; 1: Mammalia: Arctocephalus pusillus doriferus, 2:

Mammalia: Pagophilus groenlandicus, 3: Mammalia: Tursiops truncatus, 4: Aves:

Aptenodytes patagonicus, 5: Aves: Stercorarius sp. 6: Teleosti: Sardinops sagax, 7: Teleosti:

Trachurus novaezelandiae, 8: Echinodermata: Centrostephanus rodgersii, 9: Echinodermata:

Crinoidea sp. 10: Mollusca: Nototodarus sp. 11: Mollusca: Nototodarus gouldi 12: Mollusca:

Octopoda sp. 13: Mollusca: Pteriomorpha sp. 14: Crustacea: Amphipoda sp. 15: Crustacea:

Penaeidae sp. 16: Crustacea: Thysanoessa macrura, 17 Insecta: Lepidoptera sp. 1, 18:

Insecta: Lepidoptera sp. 2. 19: PCR no template control........................................................ 27

Figure 2.3 Comparison of vertebrate (top graph) and invertebrate (bottom graph) within

species K2P distance (x-axis) and between species K2P distance (y-axis) in taxa from Table

2.3. Diagonal represents equal within and between species K2P distances. Points falling on

or below diagonal line represent species that have equal or higher intra-specific divergence

compared to the pairwise divergence with at least one other species. In such cases specific

identification from sequence data based on distance measures would be ambiguous......... 29

Figure 2.4. Example of SSCP gel used to identify identical clones from a clone library. M =

molecular size marker, NTC = No template control of PCR, E3, F3… = Different clones by well

position in 96 well plate. V1, V2… = Variant clones (i.e. All V1 clones are identical). Clone

identities confirmed by sequencing: V1: Sardinella lemuru/Sardinops sagax, V2: Scomber

australasicus, V3: Trachurus novaezelandiae, V4: Chimera of Trachurus novaezelandiae and

Sardinella lemuru/Sardinops sagax sequences, V5: Tursiops truncatus, V6: Sardinella

lemuru/Sardinops sagax, 2 substitutions difference from V1, V7: Arripis georgianus, V8:

Scomber australasicus 1 substitution difference from V2, V9: Poor sequence – discarded,

V10: Sillago robusta ................................................................................................................ 30

Figure 3.1 Frequency histogram of approximate SWBD scat volumes. The numbers of final

samples used differ from the totals since the original DNA extract may have been cross-

contaminated during DNA extraction procedures and there was no sample remaining to

perform a second DNA extraction .......................................................................................... 42

Figure 3.2 Priming regions of Mugil cephalus species specific primers designed in this study.

Dots represent same nucleotide as Mugil cephalus for that nucleotide position, ~ represents

the gap between priming regions. Note the mismatch between the PCR primers (particularly

at the 3’ end of the primers) for non-target species. ............................................................. 44

Figure 3.3 Bland-Altmann plot of PCR threshold cycle (c(t)) differences between DNA

extraction kits. Bold middle line represents the mean difference between MoBio and

QIAGEN kits (MoBio c(t) – QIAGEN c(t)) and the dotted line the standard deviation of the

mean. ...................................................................................................................................... 48

Figure 3.4 Proportion of predator origin (mtDNA and NuMT) clones from different sample

volume categories of SWBD samples. Sample volume categories and number assayed

indicated on the x-axis. Filled circles are individual data points, the open circles the mean of

each volume category, whiskers bootstrapped 95% binomial confidence interval for the

mean for grouped data. .......................................................................................................... 51

Figure 3.5 Actual Proportion of prey in SWBD diet (Dark Grey horizontal lines) compared to

proportions of amplicons in clone libraries calculated by different means. 1) Prey

proportions averaged across individual scat samples (circles), 2) Prey proportions averaged

across individual dolphins (squares), and 3) absolute total proportion (triangles) by pooling

all prey amplicons in clone libraries. Whiskers represent the bootstrap 2.5 and 97.5

percentiles for 95% confidence intervals (light grey lines) and 95% confidence intervals for

the mean (red lines) due to multinomial sampling error. Symbols represent arithmetic mean

between groups (circles and squares) or total prey proportion (triangle). For comparison,

arithmetic 95% confidence intervals of arithmetic mean proportions of raw data are given

(dotted lines). Prey species as follows: Sa: Scomber australasicus, Ag: Arripis georgianus, Mc:

Mugil cephalus, N: Nototodarus spp., P: Penaeidae spp ........................................................ 53

Figure 3.6 Samples that tested positive for Mugil cephalus DNA per day of the feeding trial

(sequential days from left to right). For the first day, only samples collected after the first

positive sample were included in this figure. Excluding the first test day, numbers above x-

axis are total sample numbers for that day. Data are black dots with actual value to the right

of the data point. Data are joined by lowess smoothed line. ................................................. 55

Figure 3.7 Boxplot with overlaid probability density function estimated with kernel density

estimates of proportion of positive samples from bootstrap replicates (n = 1000) from first

to last prey detection. Filled red line is the actual proportion of positive samples over the

duration of the prey detection period; dotted red lines are the bootstrap 95% confidence

intervals of the proportion of positive samples calculated from the bootstrap variance

estimate. ................................................................................................................................. 66

Figure 3.8 Average taxon (top panel: species; bottom panel: family) amplicons per clone

library. Filled circles represent arithmetic average across individuals, open circles represent

proportion of pooled total prey amplicons. Whiskers are 95% confidence intervals for 1)

population arithmetic mean (dotted lines of filled circles); 2) bootstrapped 95% confidence

intervals for mean due to sampling error in each clone library (red line of filled circles) and 3)

bootstrapped 95% confidence intervals for proportion of pooled total prey amplicons due to

sampling error in each clone library (filled lines of open circles) ........................................... 66

Figure 3.9 Absolute frequency of occurrence (sensu Wright, 2010) of six most common prey

species in Barros & Wells (1998) compared to this study. Whiskers represent exact 95%

binomial confidence intervals. ................................................................................................ 67

Figure 3.10 Average numerical proportion of prey items across all samples (other studies)

and average proportion of prey item clones across all sample (this study). Whiskers

represent truncated 95% confidence intervals of arithmetic mean. Variance was not

available for Berens et al. (2010). ........................................................................................... 67

Figure 4.1 Characteristics of putative NUMTs recovered from all samples (A) Frequency

histogram of the proportion of putative NUMTs in the clone library from each sample (n =

15); first category from 0-0.1 includes libraries with no pNUMTs. (B) Relationship between

the proportion of putative NUMTs in the library (x-axis) and the diversity of prey species

identifiable (filled circles) and putative NUMT haplotypes (open squares) (y-axis). These

relationships were not statistically tested as the variables are not independent. (C)

Relationship between the threshold PCR cycle of the 16S PCR used to amplify prey DNA (x-

axis) and the diversity of the prey species identified (y-axis) (Kendell Tau correlation; tau = -

0.45, z = -2.24, p = 0.03) (D) Relationship between the threshold PCR cycle of the 16S PCR

used to amplify prey DNA (x-axis) and the proportion of putative NUMTs in libraries (y-axis)

(Kendell Tau correlation; tau = 0.44, z = 2.29, p = 0.02) ......................................................... 80

Figure 4.2 Minimum Evolution phylogenetic tree displaying the relationship between

putative NUMT haplotypes and cetacean mtDNA. Other laurasiatherian mtDNA was used for

an outgroup (bottom seven branches). Topology was tested by bootstrapping with 1000

replications and the consensus tree is shown. Only values at nodes with a bootstrap score of

>50% are shown. Note the grouping of all putative NUMT haplotypes within the major

cetacean clade in relation to the outgroup but still distal to and highly diverged from the

majority of cetaceans. GenBank accession numbers are displayed after species name ........ 84

Figure 5.1 Initial terminal restriction fragment experiment on Tursiops truncatus DNA from

multiple individuals. Different individuals indicated by sample identifier at top of image.

Samples with * indicate samples of the same individual collected different years. Lines

indicate problems with sample seepage from bottom of gel well. Dark lines in outer lanes

represent Lambda Hind III size markers (top marker concatenation of 23130bp & 4361bp

fragments (27691bp)) with lines drawn between the same size markers on either side of gel

to account for uneven fragment migration and/or blotting distortion across gel. ................ 99

Figure 5.2 Bal 31 time series digestion of DNA from a single individual to characterize

telomere sequence distribution in bottlenose dolphins. 1. Gel image of experiment. DNA

size in base pairs indicated down left side. Lanes; 1-7 represent 1/6 volume of total Bal 31

treatment - 0 (control), 10, 20, 30, 40, 60 and 120 minutes of Bal 31 digestion. Lanes 8-14

represent remaining 5/6 volume from lanes 1-7 further treated with restriction enzymes as

per a conventional terminal restriction fragment assay. 2. Density plot relative copy number

in each DNA size class from restriction digestion treated DNA from the same experiment.

Digest times represented in top left of graph. Note the rapid digestion of DNA in the ≈23-4

kb portion after 60 minutes of digestion; this fraction represents true telomeric DNA

(Represented by a in both panels). Note also the resistance to Bal 31 digestion of laddering

in smaller size classes and of DNA up to ≈ 8.5kb that overlaps with true telomeric DNA

(represented by b in both panels). ....................................................................................... 100

Figure 5.3 Boxplot of telomere sequence per genome values from bottlenose dolphins.

Bottom whisker is 25th percentile, box is 25th-75th percentile with median dark line, top

whisker above 75th percentile. Outliers (open circles) are defined as over 1.5x the

interquartile range. Note the two extreme outliers, both from historical samples from 1995.

.............................................................................................................................................. 101

Figure 5.4 Telomere sequence per genome values of longitudinal samples from bottlenose

dolphin individuals collected at different ages. Whiskers for data points represent telomere

sequence per genome value standard deviation.................................................................. 102

Figure 5.5 Cross sectional telomere sequence per genome values for Sarasota Bay

bottlenose dolphin samples. Where two data points are from the same individual refer to

the legend on the upper right. There was no significant relationship between telomere

sequence per genome and age with this data. ..................................................................... 103

Figure 5.6 Boxplot of telomere values obtained by densitometric analysis of initial TRF

assays performed in Fig. 5.1. Bottom whisker is 25th percentile, box is 25th-75th percentile

with median dark line, top whisker above 75th percentile. Note the lack of variation across

values. ................................................................................................................................... 104

Figure 5.7 Densitometric profiles from representative lanes from Fig. 5.1. Legend show

representative lane from samples, age of samples in parentheses. FB19 (49 years old) had

the highest telomere metric values and FB157 (An adult of unknown age) had the lowest

telomere metric value. .......................................................................................................... 104

Figure 5.8 Telomere sequence per genome values from 12 samples from 6 cow-calf pairs.

Cows are grey bars, calfs white bars, whiskers are standard error. ..................................... 105

Figure 5.9 Example of Harp Seal TRF autoradiograph. ......................................................... 106

Figure 5.10 Plot of harp seal telomere sequence per genome estimates and tooth age. There

was no significant relationship between the two variables. Whiskers represent standard

deviation of CV-filtered telomere metrics. ........................................................................... 107

Figure 5.11 Plot of harp seal telomere metrics gained from densitometric analysis of TRF

assays and tooth age. There was no significant relationship between the variables with all

data, however if the two points from 1 year old males are excluded there is a significant

relationship between the variables. There was no valid reason to exclude these data points.

............................................................................................................................................... 107

List of Tables

Table 2.1Proportions of species that contain the Pac I restriction site within the amplicon

region out of species represented on GenBank (as of September 2006). .............................. 24

Table 2.2 Forward primers tested for 3’ complementarity with sequences provided in

supplement 2. Primers 1-10 are present in the degenerate forward primer presented in the

main manuscript and used in the study and primer pair empirical trial. Primers 11-19 are not

represented within the degenerate forward primer in the main manuscript; nucleotide

positions where these primers differ from those used in the study are shown in italicised,

bold and larger font. ............................................................................................................... 26

Table 2.3 Summary of comparisons of within species divergence with between species

divergence of species within each familial (or closest higher taxon) group for 16S mtDNA

target fragment. ...................................................................................................................... 28

Table 2.4 Prey and mammal clones identified in clone libraries from 5 captive feeding trial

samples subject to Treatment 2 (Pac I digestion for predator DNA exclusion prior to PCR and

also prior to cloning). Prey species abbreviations are shown below. ). Frequency of prey

occurance across scats is presented in the ‘FOC’ column. ..................................................... 32

Table 2.5 Prey operational taxonomic units (OTUs) identified from scat samples obtained

from free-ranging Sarasota Bay T. truncatus (n = 5 different individuals). Frequency of prey

occurance across scats is presented in the ‘FOC’ column. ..................................................... 33

Table 3.1. Summary of SWBD scat samples available for assays after DNA extraction and

contamination tests. Numbers in parentheses are the actual number of samples collected in

that day. Difference between samples used and numbers in parentheses are due to material

being unavailable for a second round of DNA extraction following contamination tests...... 41

Table 3.2 Summary of SWBD scat samples available for assays after DNA extraction and

contamination tests ................................................................................................................ 43

Table 3.3 Raw data from universal prey detection assays of 12 SWBD samples. Samples with

indicate this sample was assayed as part of the first chapter and the data are included

here. Dashed lines indicate different days of the feeding trial. ............................................. 50

Table 3.4 Nuclear mitochondrial pseudogenes (NuMts) variants isolated from SWBD clone

libraries ................................................................................................................................... 51

Table 3.5 Frequency of Occurrence (FOC) scores weighted by proportion of total prey clones

for SWBD samples. Prey groups arranged in order from most to least important according to

weighted FOC scores ............................................................................................................... 52

Table 3.6 Summary of OTUs identified from Sarasota Bay T. truncatus faeces and gastric

samples using a Threshold/Topological Neighbour Joining Analysis ...................................... 58

Table 3.7 Summary of Sarasota Bay T. truncatus number of clones per individual

taxonomically assigned to prey MOTUs using SAP software and accounting for Taq DNA

polymerase error in clone libraries. If Genus and Family level identifications were > 95%

similar to another identified taxon, they are listed below the most similar (% identity)

species level identification and indented. Blue text indicated a difference in prey assignment

between the ‘% Threshold/Neighbour joining’ method and the MOTU/SAP method. .......... 62

Table 3.8 Species and family Frequency of Occurrence (FOC) and FOC scores for taxa

appearing in more than one sample. Taxa are grouped in descending order of weighted

frequency of occurrence score................................................................................................ 65

Table 4.1 Summary of the occurrence between samples of all recovered putative NUMT

haplotypes and their BLAST closest matches .......................................................................... 79

Table 4.2 Number of pairwise nucleotide differences (bottom diagonal) and pairwise genetic

distances as estimated by the Kimura 2 parameter substitution model (top diagonal)

between suspected NUMT sequences, T. truncatus and the closest BLAST match of the

suspected NUMT’s. Light grey shaded areas are the pairwise comparisons between

suspected NUMTs and true cetacean mtDNA, dark grey shaded areas are the lowest genetic

distance estimate(s) of each suspected NUMT. ...................................................................... 83

Table 4.3 Results of referencing spurious sequences obtained from fecal samples in this

study against Tursiops truncatus draft whole genome shotgun sequences on GenBank by the

BLAST algorithm. Grey shading denotes homologous matches of ≥ 98%. ............................. 85

Table 5.1 Summary of Sarasota Bay samples for telomere assays. ........................................ 92

Table 5.2 PCR assay parameters for q-PCR telomere assays. ................................................. 98

Table 6.1 Summary of methods available to study the diet of live cetaceans.....................119

1

Acknowledgements

I thank my supervisors Nick Gales, Simon Jarman and Mark Hindell. Nick for finding a place

for me in his research program, opening up doors that would usually remain closed to me

and offering advice for developing my still appalling diplomatic skills. Thanks to Simon for

offering jovial advice on many things molecular and science writing, and both Simon and

Nick for facilitating my research through their institutional funding. Thanks to Mark for

providing timely and insightful academic guidance throughout the duration of my

candidature as well as a place in his lab at UTAS (which, importantly, has a fridge with beer

in it most of the time).

Funding for this project was provided by the Australian Antarctic Division, the Winifred

Violet Estate Trust Fund, the ANZ Holsworth Wildlife Research Endowment, The Norwegian

Polar Institute, The University of Tasmania and the Australian Centre for Applied Marine

Mammal Science (now the Australian Marine Mammal Centre). Further support was also

provided by Mote Marine Laboratories and the Chicago Zoological Society. I am extremely

grateful to these organisations for their financial support of my research.

I am grateful to Randy Wells, Nelio Barros (deceased), Dammon and Janet Gannon, Jason

Allen, Aaron Barleycorn, Brian Balmer and all staff and volunteers of the Sarasota Dolphin

Research Program. The work on this thesis on Sarasota dolphins would not have been

possible without these people. I especially thank Randy Wells for being generous in his

support and facilitation of my research. Thank you to Ruth Casper who performed the

dolphin feeding trial that some of the work in this thesis was based upon and who allowed

me to use the samples. Thanks to Sarah Laverick for helping Ruth with this feeding trial and

also providing administrative support. Thanks to C. Scott Baker for help in facilitating the

small Southern Right Whale component in this thesis. Thanks also to Christian Lydersen for

facilitating my participation in the harp seal study.

Lastly thank you to my beautiful girl Krista for putting up with me (for the most part)

throughout this process.

2

Abstract

Traditional molecular ecology has focused on describing the historic processes that lead to

contemporary patterns of diversity within populations, species and higher taxa. Molecular

tools can also identify the origin of biological material and shed light on contemporary

ecological processes of populations and individuals. This thesis is concerned with evaluating

the efficacy of some of these latter nascent applications for diagnosing near real-time

ecological information in marine mammals. The applications under investigation were (i)

DNA-based methods to identify the prey of cetaceans and (ii) using the size of telomeres

within the life of an individual to estimate the age of individuals in the Pinnipedia and

Cetacea.

Diet samples like faeces are complex mixtures of predator, prey and symbiont DNA and as

such they require techniques that can exclusively target prey DNA. Previous DNA-based diet

studies had employed species- or group-specific polymerase chain reaction (PCR) primers to

achieve this and thus were implemented ad hoc or required a priori diet knowledge, which

limited their scope. I developed a prey detection method that employed novel PCR primers

widely complementary (‘quasi-universal’) for most animal 16S mitochondrial DNA (mtDNA)

and a restriction enzyme to selectively exclude predator mtDNA. The method requires no a

priori diet knowledge and can be applied to other predators with a minimum of modification.

Faecal samples were collected from two sources; captive bottlenose dolphins (Tursiops sp.)

fed a known diet and free-ranging bottlenose dolphins from Sarasota Bay, Florida.

Two techniques were applied to detect prey DNA in the captive samples; amplification of a

small mtDNA fragment using a species-specific PCR primer pairs designed to detect a known

prey species and the ‘quasi-universal’ method. Using the species-specific method, a prey

signal was detected within 4-7 hours of feeding the captive dolphins the known diet and

persisted for at least 12-19 hours after the diet ceased. After the first detection, 60 +/- 12%

(mean +/- 95% CI) of captive samples contained a prey DNA signal using species-specific

methods. The ‘quasi-universal’ method was applied to 12 samples from within the time

period with a known diet of 10 prey species from 3 Phyla. Up to six prey species were

detected per sample (range 0-6, mean 3.2 +_ 1.7 (SD) species) and all but one prey species

consisting of 2% wet weight of the total diet were detected across all samples. No prey DNA

was detected from one captive sample using this method. Estimates of prey item amplicon

amounts showed congruence with the total proportion of wet weight of most prey items in

the diet, though variability introduced through sampling amplicon clone libraries puts wide

confidence intervals on these results.

The ‘quasi universal’ method was then applied to 15 faecal and 9 gastric samples from 19

free-ranging Sarasota Bay dolphins. Thirty two prey molecular operational taxonomic units

(MOTUs) were identified across all samples (range 0-9, mean 3.7 +/- 2.2 per individual),

consisting of 28 taxonomic assignments, 18 of which were species level identifications. One

3

sample did not contain prey DNA. The difference in results between samples from the

captive animals and free-ranging animals suggest that factors such as sample collection

methods, sample amount, sample storage duration and whether animals consume live or

dead prey may affect the efficacy of DNA-based techniques, which has ramifications for

interpreting results from captive feeding trials. These results were also congruent with diet

data from this population via traditional hard-parts analysis of stomach contents from

stranded individuals.

An unexpected consequence of using restriction enzymes to exclude predator mtDNA was

the appearance of nuclear mitochondrial pseudogenes (NUMTs) in samples. The appearance

of NUMTs in 15 faecal samples from Sarasota Bay dolphins was further investigated, in

order to understand their impact on DNA-based dietary analysis in a field situation. Nine

unique NUMT paralogs detected in 13 of 15 samples were represented by 1-5 paralogs per

sample and were estimated to be between 5-100% of all amplicons produced per sample.

The diversity of prey DNA and the proportion of NUMT amplicons per sample were related

to real-time PCR cycling characteristics, with lower prey diversity and a higher proportion of

NUMTs recovered with increasing real-time PCR threshold cycle values. This indicated that

low DNA yields from diet samples are more likely to have NUMTs detected and less likely to

contain prey DNA using this technique. This predator-prey system is relatively well sampled,

which facilitated ease of identification of NUMTs, however for many study systems this may

not be the case.

Telomeres are nucleoprotein structures on the end of eukaryote chromosomes that consist

of regions containing ‘telomere-like’ and ‘true’ telomere tandomly repeated DNA sequence

and single stranded telomere sequence overhangs on the 3’end of each anti-parallel DNA

molecule, each with associated proteins. They change size throughout the life of many

animals, suggesting that they be a molecular means to estimate animal age. To examine

whether telomeres would be useful to estimate the age of pinnipeds and cetaceans,

samples were collected from populations of three model species where the age of

individuals was known or could be relatively inferred; harp seals (Pagophilus groenlandicus),

bottlenose dolphins and southern right whales (Eubalaena australis).

In Harp Seals, telomeres were measured using two techniques, (i) de-naturing terminal

restriction fragment analysis and (ii) quantitative PCR (Q-PCR). There was no relationship

between age and telomere length using either telomere measurement method, however

there was a strong correlation between the methods, indicating that they were comparable.

Telomere dynamics in cetaceans were then investigated. Previous studies had shown that

satellite DNA in Mysticete cetaceans contains telomere sequence repeats that may bias

telomere measurement techniques. This was investigated in Odontocete cetaceans by

characterizing interstitial telomere sequence (ITS – that is, telomere repeat DNA sequence

that is not a part of true telomeres) in Bottlenose Dolphins. It was found that substantial ITS

exists in bottlenose dolphins, and given its presence in closely related Mysticete cetaceans,

most likely all cetaceans. The presence of this ITS made denaturing TRF analysis difficult to

4

interpret and so attempts were made to measure Bottlenose Dolphin telomeres using non-

denaturing TRF assays and Q-PCR assays. Attempts at non-denaturing assays were

unsuccessful and further efforts focused on Q-PCR assays. No relationship was found

between Q-PCR telomere metrics and age in bottlenose dolphins. In four instances

longitudinal samples were available from the same bottlenose dolphin individual a number

of years apart and these were compared. While two samples showed the typical pattern of

decline in telomere size with age, one showed no discernable change with age and one

individual displayed an apparent gain in telomere sequence with age.

Given these results a small subsection of 6 cow-calf pair Southern Right Whale samples

were initially analysed. Four adults appeared to contain shorter telomere sequences than

their calves although in two cases, calves contained less telomere sequence then their

presumed mother. In light of these results the use of telomeres to estimate age of individual

marine mammals did not appear a valid technique and could not be recommended.

Overall, this study achieved its aims of appraising the efficacy of these nascent molecular

ecology techniques in marine mammals. In the first instance, DNA-based diet analysis

appears to hold great promise for analysis of cetacean diet. The methods were sensitive,

identified prey to the lowest taxonomic level in many cases, and made use of samples that

cannot be used for traditional diet analyses. They allow high-resolution prey detection of

live animals, a feat that cannot be otherwise achieved in cetaceans failing direct observation

of feeding events. Additionally the ‘quasi-universal’ method can be applied to any predator

where its 16S mtDNA sequence is known or can be gained, and the biases of the technique

can be inferred by using current data from databases such as GenBank. Conversely,

telomeres appear to hold little use for age estimation in marine mammals. All

methodological issues aside, there are many exogenous and endogenous influences other

than chronological time on individual telomere dynamics and these are not well understood.

It is not recommended that telomeres be used for age estimation in any animal group

without considerably more work. The latter outcome is as useful as the former, since any

emerging technique (no matter how promising) must be put through rigorous critical

appraisal in order to understand whether the application is warranted at all, or if so what

the caveats might be.

5

6

7

- Introduction Chapter 1

This thesis investigates new approaches in two fundamental aspects of the natural history,

biology and ecology of any animal species; diet and age estimation. Knowledge of both the

diet and the age of animals have been recognized as fundamental biological factors since

some of the very first scholarly observations of animals (Arisotle ≈350BC). The reasons for

this are easy to understand - in the animal kingdom everything that lives, must eat, and also

inevitably grows old. These are at first, two seemingly disparate areas of a species’ biology.

However, the uniting focus of this thesis is not the inherent biological link between them.

Rather it is the relatively recent conceptual and technological advances in molecular biology,

producing molecular techniques that can be applied to each discipline. The application of

molecular biology to these areas could potentially allow information to be gathered that is

impossible to obtain by other means.

This introductory chapter outlines each of these two areas of study, provides the

background and rationale for the work performed, and ends by outlining and justifying the

structure of the remainder of the thesis.

1.1 Diet Investigations

The struggle for resources is an underlying driver of evolution (Darwin 2003). Food is a

principal key resource, as it is the basis of energy input for animals. Energy is the overriding

currency of all life. Therefore, the flow of energy into and out of systems, whether at the

level of single cell, organism, population, or entire ecosystem, is central to understanding

and integration of all these levels of biological phenomena.

The acquisition of nourishment is strongly influential on the growth, health, reproduction

and survival of individuals. Thus, a large proportion of the activity budgets of most animals

are devoted towards procuring food (Stephens and Krebs 1986). An understanding of the

basic mechanisms of how this is performed, such as what is eaten and how this is gathered,

forms a fundamental basis to quantifying the fitness of individuals, the natural history,

ecology and life history of species and, in due course, the dynamics of ecosystems. These

areas are important in the context of basic research and biological knowledge. They are also

increasingly important in an applied sense, as humans strive to understand and manage the

impact of their activities on natural systems.

The methods used to investigate the diet of animals are diverse, however they can be

broadly divided into two categories: Direct and indirect observation. Direct observation

entails observing what an animal eats in the wild. Indirect observations involve inference of

diet through the use of some sort of indicator or marker, which informs as to what the diet

8

may be. Markers used for indirect studies may be chemical or physical, but in all cases a link

between the marker and potential source(s) of the marker is required. For wild animal

studies both methods have been employed, however indirect methods are most often used.

This is simply because it is rarely possible to observe wild animals feeding in an unbiased

fashion or even at all. For the group of animals that is the focus of this thesis, marine

mammals, this is accentuated further since most are difficult to observe at all.

Except for rare observations of foraging, marine mammal diet has always been investigated

through indirect means. Traditionally, this has involved the quantification of prey physical

remains in the faeces or stomach contents of individuals (Pierce and Boyle 1991). Over the

last three decades, marine mammal diet investigation approaches have expanded to include

chemical markers, such as biological molecules and the stable isotopes of elements. Each of

these techniques has unique characteristics and constraints in terms of the accuracy and

precision of diet information obtainable and the temporal scale over which the information

is relevant. Detection of prey remains in faeces or stomach contents is conceptually simple

and can facilitate a high taxonomic resolution of prey over short timescales (days). Yet it is

biased by the differential survival of the diagnostic remains of different prey species (Pierce

and Boyle 1991). Some biological molecules such as the fatty acids of lipids can offer a

variable taxonomic resolution of prey integrated into the predator over short to medium

(days to weeks) timescales (Budge et al. 2006). However in most cases prey can only be

assigned to broad taxonomic categories qualitatively. Stable isotopes can inform as to the

environmental conditions and diet of a consumer at the time of their integration into tissue,

over timescales of days to years, dependent on the tissue analysed (Newsome et al. 2010).

Yet, in most instances it is difficult to derive any taxonomic information on the prey

consumed with stable isotope techniques, without some sort of in situ comparison.

A relatively recent advance in studying the diet of animals has come from detecting food

DNA in faeces or stomach contents (Kohn and Wayne 1997; Symondson 2002b). This

technique is conceptually similar to detecting prey hard parts, except food DNA is detected

using the tools of molecular biology, instead of visual identification of prey hard parts. DNA-

based diet techniques are appealing because they can be used in situations where the

application of previous techniques is not advantageous, desirable, or even possible

(Symondson 2002a). For example, DNA-based techniques do not suffer from the constraints

of differential hard-parts survival, in that there is no inherent bias towards detecting prey

with more robust hard parts. For animals that are fluid feeders, such as insects, or for other

animals which digest food thoroughly, such as cetaceans and seabirds, it is not possible to

perform faecal hard-parts analysis with any reliability at all.

DNA-based diet techniques have been applied to various marine mammal species with

varying degrees of success. The proof of concept was first demonstrated using faecal

samples from cetaceans (Jarman et al. 2002), however these techniques have since been

predominately applied to pinnipeds (Purcell et al. 2004; Deagle et al. 2005b; Casper et al.

2007a; Casper et al. 2007b; Deagle et al. 2009; Tollit et al. 2009). Experimental feeding trials

9

have demonstrated high rates of detection of known prey seem to vary depending on the

predator species, although detection rates are generally superior to hard parts analysis

(Deagle et al. 2005b; Casper et al. 2007a). As well as facilitating comparatively high rates of

prey detection, these techniques have identified prey taxa unable to be discriminated with

hard parts analysis (Tollit et al. 2009). There also appears to be some quantitative signal of

the mass proportions of prey consumed (Deagle et al. 2005b; Deagle and Tollit 2006; Bowles

et al. 2011), although comparison of these types of studies reveals inconsistencies in these

findings at present (e.g. Casper et al. 2007a; Deagle et al. 2010). Despite this, application of

DNA-based diet techniques to pinniped taxa has already revealed novel trophic insights that

were not possible to detect with other methods (e.g. Tollit et al. 2009).

Compared with pinniped studies, there have been comparatively few studies on the efficacy

of DNA-based diet techniques to investigate cetacean diet. Early studies demonstrated that

prey DNA could be detected in mysticete faeces (Jarman et al. 2002; Jarman et al. 2004).

These studies were important because they showed that there was potentially a technique

that could yield specific prey data from cetacean faecal samples, which was previously not

possible. Most other cetacean diet studies either did not yield specific prey information (e.g.

stable isotope or fatty acids studies), or those that did were performed by visual

examination of the stomach contents of stranded individuals, or those that were by-catch

from commercial fishing operations. There has long been a doubt as to whether data from

stranded or by-caught individuals was representative of the wider population, and few ways

to test the assumption that they were (Barros and Clarke 2002). Furthermore, DNA-based

diet techniques had never been applied to any odontocete cetaceans.

This thesis attempted to address some of the basic outstanding issues when applying DNA-

based diet techniques to cetaceans, using bottlenose dolphins (Tursiops sp.) as a model

species. For example, are these techniques amenable to studying odontocete diet? If so,

does sample quality and/or quantity effect this? How long after ingestion can a prey species

be detected and what is the temporal persistence of the DNA signal from consumed prey?

How well do DNA-based methods describe the diet qualitative and quantitatively, if at all?

Could these techniques be used to test some of the assumptions of traditional approaches

such as hard parts analysis? What are the limitations of these techniques when applying

them in unknown situations, such as studies of wild animals? These were the questions that

the DNA-based diet component of this thesis aimed to address.

Methods were developed to detect diverse prey DNA templates from predators with

unknown diet and the efficacy of these was examined using captive animals fed a known

diet. The techniques were then applied to samples from a wild free-ranging population to

examine their diet and test some assumptions of previous research findings. Finally, some of

the technical complexities of using DNA as a dietary marker were demonstrated and

recommendations on interpretation of results are provided that are applicable to all DNA-

based diet studies.

10

1.2 Age-based Investigations

‘Getting old’ is familiar to us all and is almost a ubiquitous phenomenon in the animal

kingdom. Biologically, ‘getting old’ or ageing, is manifest through the progressive physical,

biochemical , physiological and behavioural changes occurring during the processes of

development, senescence and finally death. These changes are continuous and generally

more pronounced the further apart they are in the lifetime of an individual organism, but

not necessarily in a linear manner. The age of an organism is therefore a time-referenced

record of the state of an organism along its life-schedule, relative to its birth. Thus, an

understanding of the age related changes in an organism, and how long it lives, are crucial

to understanding the life history and biology of the organism itself (Caughley 1977; Stearns

1992). Furthermore, by understanding the changes through time that individuals experience

and how this varies between individuals, it can be ascertained how these dictate population

dynamics processes. Although of fundamental theoretical importance, knowledge of age

related change and population dynamics processes of organisms is critical to detecting,

understanding and managing anthropogenic and natural impacts on individuals and

populations (e.g. Holmes and York 2003).

With the recognition that knowledge of age is fundamental to studying organisms and

populations, there is an obvious need for ways to estimate the age of study subjects. As

with diet methods, age estimation methods can be broken down into the categories of

direct and indirect methods. Direct age estimation methods estimate the age of individuals

through the use of some intrinsic marker which is a record of chronological time. Indirect

methods employ markers that assess the state of development of one individual against the

developmental state of others. Both methods have been widely employed for the purpose

of age estimation, but direct methods are preferred since they are less prone to being

confounded by unpredictable biological variation in the age marker employed.

Age estimation methods are highly diverse. Just about any variable that varies predictably

with age has been used to estimate the age of animals. For example, the diameter of

elephant faeces has been used to estimate the age of the elephant that produced the faeces

(Reilly 2002; Morrison et al. 2005). Similarly, the accumulation of particular biological

molecules with increasing age has been used to estimate the age of invertebrates (Ju et al.

1999). Both of these quite different methods are indirect, since they require the

relationships between the marker and age to be established in known age animals, before

they can be applied to unknown individuals. Examples of direct age estimation methods can

be found amongst studies of vertebrates, where hard tissues are deposited periodically with

respect to chronological time in some taxa. Examples of this include periodic deposition of

lamina or ‘growth rings’ in the otoliths of some fish species (Campana 2001), or of dentine

or cementum in the teeth of mammals (e.g. Hohn et al. 1989; Childerhouse et al. 2004).

There are also some aspects of bone growth and maintenance that vary with chronological

11

time in predictable ways and these have been used to estimate the age of some reptiles and

mammals (Zug et al. 1986; Marmontel et al. 1996).

For many animals there are no reliable techniques to estimate age and most of the

techniques with highest accuracy require either destructive sampling to access the tissue, or

highly invasive capture, restraint and surgical procedures to extract the tissue. These can be

justified in many instances, however where the study species is endangered, or where the

procedures may impact animal welfare significantly or affect the quality of subsequent data,

such procedures are highly questionable and often contentious (e.g. Fiesta-Bianchet et al.

2002 and Nelson 2002; Gales et al. 2005). If they can be applied at all, these techniques

become progressively less desirable the rarer, larger, and more long-lived a taxa is, because

these taxa are less resilient to impacts than short-lived taxa (Western 1979). Ironically, it is

also these taxa where age information is most important for understanding population

dynamics processes, since population processes are highly age structured in long-lived taxa.

Many marine mammals are large, long-lived taxa where gaining age information for

individuals is difficult. Most reliable marine mammal age estimation techniques require

highly invasive or destructive sampling, as they are based on counting periodic increments

in teeth or bone, or for baleen whales, timpanic bullae. There are also techniques which

measure the state of racemisation of asparctic acids enantiomers which require destructive

sampling. Recently, there has been application of bubber fatty acid profiles which vary with

age predictably to estimate the age of humpback and killer whales (Herman et al. 2008;

Herman et al. 2009). This last technique is the only one available that can be applied to live

marine mammals without capture and restraint to remove teeth. However, there is still

much work in understanding the mechanisms behind the relationships and how general

they may be applied within and between taxa (Herman et al. 2008; Herman et al. 2009).

There is a wealth of information that can be gained by ascertaining the in situ age structure

of marine mammal populations and the age of individuals followed in longitudinal studies,

so there is a strong desire amongst marine mammal biologists to develop techniques that

will facilitate this. At the time that the work on this thesis began (2005) there was no

reliable method to estimate the age of live marine mammals of unknown history without

capture to remove teeth, which is not possible for many marine mammal species.

However, there was a suggestion in the literature that telomeres may be used as a genetic

marker to estimate the age of live animals (Haussmann and Vleck 2002; Vleck et al. 2003;

Nakagawa et al. 2004). Telomeres are nucleoprotein structures on the ends of all eukaryote

linear chromosomes. They consist of tandemly repeated DNA sequence ending in a folding

loop structure at the terminus of the chromosome (Greider 1999), with telomere-specific

proteins along their length (Blackburn 1991). Telomeres have many known functions

including facilitating chromosome pairing during cell replication, signally cell cycle arrest and

importantly for the purposes of age estimation, protecting coding DNA from sequence loss

during the process of DNA replication (Blackburn 1991). Sequence loss occurs during DNA

replication because of the unidirectional synthesis of DNA and the biochemical properties of

12

DNA replication molecular machinery. DNA replication is always incomplete on the lagging

strand and DNA is lost during each cell cycle. This is known as the ‘end replication problem’

and was realized theoretically, long before it was demonstrated empirically (Watson 1972;

Olovnikov 1973). One of the principle functions of telomeres is to provide ‘expendable’ DNA

sequence at the end of chromosomes to buffer the loss of coding DNA due to the end

replication problem. When particular telomeres become critically short and unable to

maintain the loop structure on chromosome ends (i.e. are ‘uncapped’; Blackburn 2000),

they signal an arrest of the cell cycle and the cell enters a senescence phase and eventually

apoptosis (Henmann et al. 2001).

The fact that there is DNA lost during each cell cycle is a problem for tissues with high

replicative schedules throughout the life of organisms, such as germ line tissue. In such

tissues there is often an enzyme active called telomerase, a true reverse transcriptase,

which can elongate telomeres de novo (Blackburn 1991). While telomerase activity is high in

germ-line tissues, it is usually inactive or has very low activity in other tissues such as most

somatic tissue. Thus, in many somatic tissues across many taxa, there is a decrease in

telomere length with age, as cellular replication maintains somatic tissue continually with

age. For high maintenance tissues with a large turnover of cells, such as blood and skin,

there is a pronounced reduction of telomere length with age in vivo (Hastie et al. 1990;

Coviello-McLaughlin and Prowse 1997; Haussmann et al. 2003b). Early studies

demonstrated that there was a significant relationship between telomere and age in many

tissues (Harley et al. 1990; Friedrich et al. 2000) and these, in combination with the biology

of cellular and DNA replication, were the basis for the suggestion that telomeres could be

used as an age marker in animals of unknown history (Haussmann and Vleck 2002). The

overriding rationale for use of telomeres as age markers is to assay a tissue with a high rate

of mitotic turnover that is continual throughout life, such that telomere changes with age

will likely be most pronounced and continual.

The age estimation work performed in this thesis examined whether telomeres would be

useful for age estimation of pinnipeds and cetaceans. Skin, and in some cases blood, was

gathered from individuals of known age, or those where age could be relatively inferred.

Three model species were selected, mostly on the basis of availability of supporting data,

which represented three phylogenetically distinct marine mammal taxa: Bottlenose

dolphins, Southern Right Whales (Eubalaena australis) and Harp seals (Pagophilus

groenlandicus). Methodologies were assessed, telomere assays were performed and the

nature of change of these data with age was assessed for each species.

13

1.3 Structure of the thesis

Chapter 2: Since the DNA-based diet studies in this thesis were examining the diet of a

known generalist predator with a diverse diet, a molecular method was developed that

could detect diverse prey DNA from a diet sample. At the time that work started on this

thesis, the prevailing strategy in DNA-based diet studies was to target specific prey species

or taxonomic groups (Symondson 2002a; Jarman et al. 2004). These methods were

insufficient for the scope of work for this thesis as they required a priori knowledge of diet.

Some of the work performed was to test the assumptions of previous diet studies on wild

cetaceans, therefore using a priori diet knowledge to develop methods would have biased

these comparisons unacceptably.

Chapter 3: This chapter applies the method developed in Chapter 2 to faeces samples

collected from captive bottlenose dolphins fed a known diet, to test the efficacy of the

method for describing predator diet qualitatively and quantitatively. A species-specific assay

was also developed and applied to these samples to examine specific prey detection rates

and DNA signal persistence. The technique from Chapter 2 was also applied to diet samples

from free-ranging bottlenose dolphins and these results, the first to describe the diet in

detail of live free-ranging cetaceans, are compared to previous diet studies from this

population based on stomach contents analysis.

Chapter 4: This chapter examines some of the complexity of applying DNA-based diet

methods for detecting diverse prey DNA, such as the one developed in Chapter 2, to

samples from wild animals. It describes some of the challenges in assigning taxonomic

identity to DNA sequences using available databases, where there are few a priori data

available regarding the molecular diversity of predator and prey taxa. The chapter

specifically examines the confounding effects that nuclear mitochondrial pseudogenes may

have on DNA-based diet studies.

Chapter 5: This chapter examines and applies telomere assay methods for telomere-based

age estimation to the model species in this study. Suitable (or the only available) telomere

assays were applied to all study species and the relationships of telomere assay data with

age are examined. Recommendations were produced regarding the use of particular

telomere assay methods in taxa where genomic elements are uncharacterized and the use

of telomeres for age estimation in these taxa.

Chapter 6: This chapter presents a broad overview, general discussion and ramifications of

results from the entire thesis.

14

Chapter Two –

A universal DNA-based diet technique for application to potentially any predator

This chapter has been removed for

copyright or proprietary reasons

Published as :

Dunshea, G. (2009). DNA-based diet analysis for any predator. PLoS One 4(4): e5252.

doi:10.1371/journal.pone.0005252.

38

39

- DNA-based diet analyses to describe cetacean diet; Chapter 3

experimental and empirical evaluations

3.1 Introduction

Knowledge of diet is fundamental to understanding the ecology of any predator

(McLeod et al. 2003), as well as broader integration in trophic ecology and food webs

(Santos et al. 2001; Boyd 2002; Iverson et al. 2004), predator evolution (Gygax 2002; Vitt

and Pianka 2005) and competition processes between species (Heithaus 2001). However,

gathering diet information for marine mammals is particularly challenging due to the

inability to observe feeding while they are submerged or far from land (Barros and Wells

1998). Such constraints necessitate indirect methods for inferring marine mammal diet

such as identification of prey remains parts in stomachs or faeces (hard parts analysis;

Clarke and Macleod 1976; Santos et al. 2001; Flinn et al. 2002), analysis of fatty acid

(Iverson et al. 2004) and/or stable isotope (Newland et al. 2009; Newsome et al. 2010)

signatures in predator tissues and detection of prey DNA in predator faeces (Jarman et

al. 2002; Casper et al. 2007b; Deagle et al. 2009). Each of these methods offers a

different level of detail regarding taxonomic resolution of prey and the timescales over

which prey were ingested. Prey hard parts and detection of prey DNA can offer species

level identification of prey but generally over short (i.e. days) timescales (Deagle et al.

2005b; Casper et al. 2007a). Stable isotope techniques offer little taxonomic detail of

prey but may indicate relative trophic position over considerable time periods (weeks,

months and years). Fatty acid signatures provide a timescale of weeks to months (Budge

et al. 2006) and can identify lower level taxa (Iverson et al. 2004), if considerable

information on prey fatty acid variation and predator metabolism are available (e.g.

Iverson et al. 2004).

Compared to other marine mammal predators, studies of cetacean diet are further

limited by the nature of, and difficulty in obtaining, samples. Stable isotope studies from

biopsies have provided valuable insight into the trophic ecology of cetacean species and

communities (Valenzuela et al. 2009; Riccialdelli et al. 2010), yet lack detail on specific

diet composition. Contemporary investigations aiming to identify cetacean prey taxa

have used hard parts analysis of stomach contents from fisheries by-catch and stranded

animals (e.g. Barros and Odell 1990; Barros and Wells 1998; Evans and Hindell 2004;

Berens McCabe et al. 2010; Jansen et al. 2010). This has yielded information on prey

identity, relative composition and prey size. However, doubt exists whether data gained

from stranding and by-catch samples is representative of the entire population (Santos

et al. 2001; Barros and Clarke 2002). Stomach contents data also have well-recognised,

but unquantified, biases such as under-representation of particular prey with soft bodies

or those lacking robust bone structures and contention regarding prey size

reconstructions (Barros and Clarke 2002). It is possible in some circumstances to collect

faecal material from live cetaceans, however cetacean digestion is thorough and faecal

material generally consists of an amorphous slurry devoid of visually diagnostic predator

remains (Jarman et al. 2002; Jarman 2004; this study).

40

DNA-based diet techniques can identify prey in samples without the requirement of

visually recognisable prey parts (Symondson 2002b; King et al. 2008) and have been

proven successful for identifying prey in cetacean faeces (Jarman et al. 2002; Jarman et

al. 2004; first chapter). So far these studies have been restricted in terms of the

molecular approaches used to detecting prey DNA and small sample sizes. Yet questions

remain regarding prey DNA detection rates and passage times, and the ability to

reconstruct diet qualitatively and quantitatively have not been explored. Furthermore,

no studies have addressed the ability of DNA-techniques to compliment other

approaches, and allow appraisal of their biases. This latter point is particularly relevant

since at present there are no other practical means of gaining specific diet information

from free-ranging individuals.

This chapter aims to address some of the gaps in knowledge regarding the use of DNA-

based techniques to study cetacean diet. Specifically, a study of captive bottlenose

dolphins (Tursiops sp.) attempted to determine (i) prey detection rates, signal

persistence in scat samples and variation of these between individual dolphins, and (ii)

how the results were able to reconstruct diet, qualitatively and quantitatively. The

approach was then applied to diet samples (faeces and gastric) from free-ranging

bottlenose dolphins (Tursiops truncatus). The diet of this population is well studied by

stomach contents analysis of stranded animals (Barros and Wells 1998; Berens McCabe

et al. 2010) and experimental evaluations of their hypothesised foraging mode (Gannon

et al. 2005). Finally, these results were compared to stomach contents from the same

population data. I also compared approaches to identifying prey taxa from the DNA

sequence data typical of the results of these types of studies, to illustrate the difficulties

and pitfalls of assigning DNA sequence data to particular taxa in the absence of

complete databases.

3.2 Methods

3.2.1 Initial Tests of Efficacy of Different Protocols for Dolphin faeces DNA

Extraction

Two commercial kits (UltraClean™ Fecal DNA Isolation Kit (MOBIO Laboratories,Inc.) and

QIAamp DNA Stool Mini Kit (QIAGEN Inc.)) were compared for efficacy of DNA extraction

for dolphin scat samples. DNA was extracted from five samples and from at least 2 days

within the ‘Test’ diet period, using both kits following manufacturers protocols (with

modifications for viral DNA extraction for the QIAGEN kit as per Chapter 2). The DNA

extracts from both kits were assayed for presence of Mugil cephalus DNA using Real-

Time PCR with PCR primers BM-1Fwd and BM-2Rv and appropriate reaction and

thermocycling conditions PCR primers (See Below). This prey species was chosen

because of initial robust PCR results from species-specific primer trials, after prey

specific PCR primers when designed and optimised. Reaction profiles were detected on

RotorGene Real Time PCR thermocycler systems (Corbett Life Sciences) and PCR

threshold cycles (c(t) values) were determined for each DNA extract from each sample.

Within each sample consistently lower c(t) values were noted from the QIAGEN DNA

41

extracts indicating a better DNA extraction efficacy with this product (in the case of

dolphin scat samples – See Results). In light of this, the QIAGEN kit was used for all

subsequent DNA extractions.

3.2.2 Contamination Control of DNA Extracts and PCR Assays

For each batch of DNA extractions, a blank DNA extract was included in the batch and in

random order of the pipetting/centrifugation/decanting procedures, to control for

cross-contamination during the DNA extraction procedures. There were up to 23 scat

DNA extractions for each blank extraction, corresponding to the microcentrifuge

capacity available. Blank DNA extracts from each batch were assayed with the universal

PCR primers presented in chapter 2 (conventional PCR for 35 cycles with agarose gel PCR

product check), and in the case of the Sea World Bottlenose Dolphin samples, the Mugil

cephalus species-specific DNA assay (since this detected a smaller DNA fragment – see

below). If any batch blank DNA extraction was found to produce an amplicon signal or

product, this batch of DNA extractions were discard and if more sample was available,

the DNA extractions performed again. Assay PCR runs included 1-4 no template controls

to monitor for cross-contamination at PCR set-up.

3.2.3 Samples Used

3.2.3.1 Seaworld Feeding Trial Experimental Diet Samples

In light of the DNA extraction trial results, the remaining samples from the Seaworld

feeding trial first ‘Test’ (See Chapter 2) period had DNA extracted as described in

Chapter 2 and were used in this chapter. DNA was extracted from samples from 2 days

before the commencement of the ‘Test’ period, through the ‘Test’ period to four days

after the final ‘Test’ period day (Table 3.1). Throughout the text, these samples are

referred to as ‘Sea World Bottlenose Dolphin (SWBD)’ samples. A summary of SWBD

samples analysed in this chapter is presented in Table 3.1.

Table 3.1. Summary of SWBD scat samples available for assays after DNA extraction and contamination tests. Numbers in parentheses are the actual number of samples collected in that day. Difference between samples used and numbers in parentheses are due to material being unavailable for a second round of DNA extraction following contamination tests.

Period Non-Test

Test Non-Test

Day 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Totals

Samples Used

1

(4) 4

(6) 12

(12) 7

(9) 12

(13) 3

(4) 4

(5) 5

(5) 8

(8) 5

(5) 9

(11) 8

(8) 8

(10) 6

(8) 7

(11) 99

(119) Female 1 2 7 3 7 3 4 3 6 4 7 4 5 4 2 62

Male 0 2 5 4 5 0 0 2 2 1 2 4 3 2 5 37

ID Certain?

1 2 9 3 8 3 4 4 6 3 4 4 5 1 6 63

42

Approximate sample volumes were allocated to four categories using container

dimensions and measuring the depth of the scat material settled below the clear

ethanol after samples had sat undisturbed. The categories were: 1) ‘Scraps’ – Less than

1mm in bottom of sample container and only very small amounts in bottom of container

– could see through >50% of the clear bottom of the sample container, 2) <1.26 cm3 –

Less than 1mm in bottom of sample container but >50% of the bottom of the container

covered, 3) 1.26 – 3.77 cm3 – Between 1mm and 3mm in the bottom of sample

container, and 4) >3.77 cm3 – More than 3mm in the bottom of the sample container.

Most SWBD samples were in the <1.26 cm3 category (Fig. 3.1).

Figure 3.1 Frequency histogram of approximate SWBD scat volumes. The numbers of final samples used differ from the totals since the original DNA extract may have been cross-contaminated during DNA extraction procedures and there was no sample remaining to perform a second DNA extraction

3.2.3.2 Sarasota Bay Free-ranging Diet Samples

Scat samples and gastric samples were collected from 19 individual Sarasota Bay

Bottlenose Dolphins (Tursiops sp.) (Table 3.2). Scat samples were collected

opportunistically directly from the animal when they defecated while being handled for

measurements and collection of other samples (Wells et al. 2004). All but three scat

samples were collected on the deck of the research vessel. In 2006, three samples were

hand scooped from the water where the dolphins were being handled. For the samples

collected from the water, a control water sample (20ml of seawater) was also collected

up-current, in the vicinity of where the scat sample was collected.

Gastric samples were collected from 9 animals (2005 n = 5, 2006 n = 4) by running a

10mm plastic hose down the oesophagus of captured animals while they were on deck

of the research vessel. The hose was passed down to the fore-stomach and gastric

juiced were drained off into a 50ml falcon tube, which was subsampled into 5ml falcon

tubes. In five cases, gastric samples were available from the same animal from which a

19

48

26

6

28

54

31

0

10

20

30

40

50

60

Scraps <1.26 cm3 1.26 – 3.77 cm3 >3.77 cm3

Total Samples

Final Samples Used

43

scat was collected, allowing a comparison of the two sample types. Approximate sample

volumes for each sample type were measured as per SWBD samples. Table 3.2

summarizes samples collected from Sarasota Bay dolphins.

Table 3.2 Summary of SWBD scat samples available for assays after DNA extraction and contamination tests

Year IDcode Faeces Gastric Sex Age Length Faecal Volume

(cm3)

2005* SBD146 M 9 247 23.1

2005* SBD155 F 15 243 68.4

2005* SBD182 M 18 263 19.8

2005 SBD195 F Adult 258 65.9

2005 SBD199 F 3? 204 3.9

2005 SBD234 M Adult 221

2005* SBD236 M Adult 242 66

2005* SBD238 M 3? 209 72.6

2006 SBD10 M 25 257

2006 SBD20 M 17 273 6.6

2006 SBD90 F 36 255 36.3

2006 SBD113 F 10 228 14.5

2006 SBD133 F 7 217 25.1

2006 SBD151 F 6 210 23.8

2006 SBD157 F Adult 265 25.1

2006 SBD164 M 17 255 26.4

2006 SBD179 F 4 209 14.5

2006 SBD240 M 2 208 42.9

2006 SBD246 M 2 182

*Data from samples presented in Chapter 2

Samples (faeces and control) were stored for the day at 4°C until they were fixed by

addition of one volume of 100% molecular grade ethanol . Faeces and water samples

were then stored at-20°C until laboratory analysis. Gastric samples were stored for the

day at 4°C and were then stored at -80°C until laboratory analysis. Throughout the text,

these samples are referred to as ‘Sarasota Bay’ samples.

3.2.4 Universal Assays with SWBD samples

Universal diet assays described in Chapter 2 were undertaken on an additional 7

samples from within the test diet period. Additional samples were selected on the basis

of different scat volumes from at least two days within the ‘Test’ period and certainty of

the individual producing the scat. Including results from chapter 2, a total of 12 samples

were assayed with this methodology. Because of the clear effect of sample volume on

this assay (see results Fig. 3.4) and the fact that most samples were small volume

samples (Fig. 3.1), no further samples were analysed with this method.

44

3.2.5 PCR Primer Design and Testing

A species-specific PCR assay was designed to detect Mugil cephalus DNA in the Seaworld

samples. Here 16S sequences from all prey species fed to the dolphins were aligned and

species specific primers were designed using Amplicon Software (Jarman 2004). Primers

were designed to have maximum mismatch between non-target prey species in the 3’

end of both PCR primers and to have otherwise good PCR primer characteristics (i.e.

length, GC content, non-complimentary etc) (Fig. 3.1). Prey sequences were available

from GenBank, or were generated as part of this study as per Chapter 2.

The Mugil cephalus species specific PCR primers were BM-1Fwd (5’-3’):

GCTTAACTGTCTCCTTTTCCCAAC, and BM-2Rv (5’-3’) ACGTGGAACAGGGTTCATTTG, for

amplifying a 162 base pair (bp) length of Mugil cephalus 16S mtDNA (Fig. 3.1). PCR

conditions were initially optimized with AmpliTaq Gold Hot Start Taq DNA polymerase

chemistry using a Taguchi method modified for PCR optimization (Cobb and Clarkson

1994). AmpliTaq Gold chemistry was used for initial amplification and specificity trials

and afterward PCR master mix chemistry (SensiMix – BioLine, Australia) was used. Final

PCR reaction conditions were 1x SensiMix, 0.4µM each primer and 2.5µl scat template

DNA in a 12.5µl reaction volume. For trials of specificity and amplification efficiency

where prey template DNA was extracted direct from prey tissue, 1µl of template DNA

(concentration ≈1-50ng/µl) was used. The final thermal profile used for this assay was

95°C for 7.5 minutes then 40 cycles of 95°C for 15s, 60°C for 20s and 72°C for 20s with a

final extension of 72°C for eight minutes. The melt curve thermal profile was 78°C for

30s followed by sequential 2s holds at subsequent steps of 0.1°C to 95°C.

The specificity of Mugil cephalus species specific PCR primers was examined empirically

by attempting to amplify DNA from 2-5 individuals of each non-target prey species and

examining amplification success with agarose gel electrophoresis initially, then with real-

time amplification and melt curve profiles. Reaction efficiency was calculated from the

slope of c(t) values from a serially diluted plasmid standard containing the Mugil

cephalus LR-J-12887 and LR-N-13398 16S mtDNA fragment.

Figure 3.2 Priming regions of Mugil cephalus species specific primers designed in this study. Dots represent same nucleotide as Mugil cephalus for that nucleotide position, ~ represents the gap between priming regions. Note the mismatch between the PCR primers (particularly at the 3’ end of the primers) for non-target species.

45

3.2.6 Real Time PCR screening for Mugil cephalus DNA in SWBD samples

SWBD samples from two days prior to four days after the first ‘test period (n=99) were

screened for the presence of Mugil cephalus DNA with real-time PCR using Rotorgene

thermocyclers (Corbett Life Sciences) and reactions were set up using the Corbett liquid

handling robotics system. Both c(t) values and melt curve profiles were used to confirm

the amplification of Mugil cephalus 16S mtDNA in samples, as in some cases non-specific

amplification of some other template occurred which resulted in amplicons with

different melt temperature profiles. All samples were run on two separate occasions and

there was no disagreement between runs. The initial intention for the SWBD samples

was to design species specific PCR primers for all ‘Test’ prey species and screen each

sample for them, however after initial results with the Mugil cephalus PCR primers and

the universal prey detection assay these plans were abandoned (See Results and

Discussion).

3.2.7 Universal Assays with Sarasota Bay samples

All additional scat and gastric samples from 2005 and all scat and gastric samples from

2006 were assayed with the universal method (Chapter 2). For these samples up to 19 -

41 clones per individual (19 - 20 per scat; 15-38 per gastric sample) were scored for

sequence variants with single stranded conformational polymorphism analysis prior to

sequencing representative variants (see chapter 2). DNA was extracted from gastric

samples as per the procedure used in Chapter 2 for scat DNA extraction, except the

whole gastric sample (i.e. 5ml falcon tubes) was spun down for 5 minutes at 12000g and

the clear gastric juice decanted. DNA extraction was performed on the solid material

that pelleted at the bottom of the centrifuge tube.

3.2.8 Data Analysis

3.2.8.2 SWBD Data Analysis

Since the SWBD samples were a sub-sample of all defecations within the time period of

the study, bootstrapping was used to estimate the reliability of the estimate of the

proportion of positive samples for Mugil cephalus DNA. There were 68 samples

(inclusive) between the first and last positive sample. The bootstrap procedure re-

sampled these data with replacement 68 times and the number of positives was

recorded. This procedure was repeated 1000 times and bootstrap sample variance used

to calculate confidence intervals for proportions estimates (Efron and Tibshirani 1993).

Where clones were considered of predator or non-predator origin, mean binomial

proportions within groups and confidence intervals (95%) were estimated using the

analysis of overdispersed data (‘aod’) package in R (Lesnoff and Lancelot 2010). The

bootstrap (1000 replicates) estimates of mean binomial proportion and 95% confidence

intervals are presented.

46

Three approaches were used to compare whether prey amplicon proportions from

SWBD universal assays reflected wet weight proportions of diet species; 1) prey

amplicon proportions were calculated for each fecal sample and averaged across all

samples, 2) prey amplicon proportions were calculated for the pooled clone libraries

from each individual dolphin and averaged across individuals, and 3) all clone libraries

were pooled and absolute proportion of prey amplicons compared to actual diet.

Proportions of prey amplicons in clone libraries had two sources of variability that

required description. The first was variability in amplicon proportions within each clone

library due solely to error from sampling a multinomial population. The second was

variation between the groups of samples (i.e. scats and individual dolphins). To estimate

the effect of multinomial sampling error on the mean proportions of prey amplicons

across scats, for individuals and for total proportions, a bootstrapping approach was

implemented. Since the sampling unit of all levels of analysis (i.e. scat, individual and

pooled total) was individual sample clone libraries, the prey amplicons of each clone

library were re-sampled with replacement with a sample equal in size to the original

sample of prey amplicons for that sample. These data were collated by scat, individual

and the total pool and proportions calculated and averaged where applicable (i.e.

between scats and individuals). This procedure was repeated 1000 times and the

bootstrap sample variance used to calculate 95% confidence intervals for the estimate(s)

of the mean or the total proportion. For across group means, the 95% confidence

intervals were also calculated for each bootstrap replicate and the 2.5 and 97.5

percentile value of these used to represent the 95% confidence interval for variability

across groups (i.e. fecal samples and individuals), which takes into account the effect of

multinomial sampling error on across group variation.

3.2.8.3. Sarasota Bay Samples Sequence Data Analysis

Two different species assignment processes were implemented for prey sequences

recovered from Sarasota Bay samples. Firstly, a “percent threshold, neighbour joining

topology” approach identical to the process described in Chapter 2. Secondly, all prey

sequences from all samples were examined and edited concurrently and primer regions

removed. Sequences were assigned into molecular operational taxonomic units (MOTUs)

on the basis of similarity allowing for Taq DNA polymerase error, since cloned amplicons

of the same sequence may differ due to Taq DNA polymerase error. Following Thalmann

et al. (2004), with a Taq DNA polymerase error rate of 7.3 x 10-5(Kobayashi et al. 1999), it

was estimated that a sequence could be the same as a reference if it contained ≤ 2

substitutions over 220bp. However, if the same MOTU containing ≤ 2 substitutions

relative to a reference was present in more than one sample, it was interpreted that the

sequence was a genuine variant since the likelihood of the same random substitution(s)

occurring at the same nucleotide position(s) in the amplicon in different PCRs is

exceedingly low. If MOTUs were not a 100% match to database reference sequences,

they were assigned to the lowest taxonomic unit using Bayesian phylogenetic analysis,

implemented in the Statistical Assignment Package (SAP) (Munch et al. 2008a). The SAP

estimates the posterior probability that a sequence is part of a monophyletic group by

comparing the sequence to databases, gathering significant homologues and associated

47

annotated taxonomic data and performing Bayesian phylogenetic analysis (Munch et al.

2008a). Posterior probabilities for monophyletic groups are estimated using Markov

chain Monte Carlo sampling of all possible Bayesian phylogenetic trees (Munch et al.

2008a) and the lowest level taxon with ≥ 95% prior probability was assigned to the

sequence. For all taxonomic assignments using either method, the assignments were

finally hand-curated to assess if the matching taxon had a distribution that included

Sarasota Bay and surrounding waters using online databases such as FISHBASE

(www.fishbase.org). In some cases further lower level taxonomic assignment, or limited

options for identity were made available at this last step, since there may be only one

genus or few species that are known from the study area.

Because chimeric DNA sequences can be obtained when PCR amplifying mixed,

degraded DNA sources, all sequences unable to be identified to at least Genus were

also analysed with a different more comprehensive chimera detection software package

than that used in Chapter 2; ChimeraSlayer (Haas et al. 2011). Although nuclear

mitochondrial pseudogenes (NUMTs) can confound prey identification using mtDNA

(See Chapter 4), sequences < 98% similar to the closest prey match were all most closely

matching to teleosts and thus were not examined for NUMT origin, given the

conspicuous absence of NUMTs in this chordate group (Bensasson et al. 2001b;

Venkatesh et al. 2006). Sequences identified as predator NUMTs were recovered from a

number of faeces samples from both SWBD and Sarasota bay dolphins.

Proportions of each prey type in the individual clone libraries were estimated based on

variant proportions in the sampled clones. Given the results of the SWBD clone library

analysis, inference regarding the importance of different prey types to Sarasota Bay

dolphins was made using a weighted frequency of occurrence approach and averaging

prey amplicon proportions across individuals. The same bootstrapping procedure was

applied to these data as was used for the SWBD clone library data to represent the

effect of multinomial sampling error on pooled amplicon totals and between individual

means and 95% confidence intervals.

When comparing prey occurrence and amount values between this and past studies,

95% confidence intervals were calculated for inference of difference between studies

following (Wright, 2010). Exact binomial confidence intervals were calculated for

frequency of occurrence values and arithmetic confidence intervals were calculated for

numerical proportions and proportion of prey amplicons across clone libraries.

48

3.3 Results

3.3.1 DNA Extraction Test

The QIAGEN kit yielded lower PCR threshold cycle (c(t) ) values for four out of five

samples positive for Mugil cephalus DNA (Fig 3.2). The MoBio kit had c(t) values on

average 3.59 +/- 2.75 (S.D) cycles higher than the QIAGEN kit (Fig 3.2), indicating a

comparatively lower DNA yield (on average ≈12 times less prey DNA given exponential

DNA amplification at perfect PCR efficiency) and thus the QIAGEN kit was used for

subsequent DNA extractions.

Figure 3.3 Bland-Altmann plot of PCR threshold cycle (c(t)) differences between DNA extraction kits. Bold middle line represents the mean difference between MoBio and QIAGEN kits (MoBio c(t) – QIAGEN c(t)) and the dotted line the standard deviation of the mean.

3.3.2 Contamination Control of DNA Extracts and PCR Assays

DNA extracts for SWBD samples were extracted in 15 batches with between 3-23

samples in each batch. There was no sign of cross-contamination when extraction blanks

from each batch were assayed with the universal PCR primers presented in Chapter 2.

However, when extraction blanks were assayed for Mugil cephalus DNA using species

specific primers, 7 of the 15 batches had late (>34 cycles) amplification signals for Mugil

cephalus DNA. All DNA extracts from these batches were discarded and DNA was

extracted again from these samples with fresh reagents. Of the 48 samples that needed

re-extraction, 19 did not have sufficient material since all material was used in the

49

original DNA extraction (see Table 3.1, Fig 3.1 for details). New DNA extract batch blanks

(n = 2) were assayed for contamination and no amplification was apparent from

universal and Mugil cephalus DNA assays. After this quality control was performed the

total number of SWBD samples available was 100 from 119 with a maximum of 4

samples lost in a single day due to contamination at the initial DNA extraction step (see

Table 3.1).

Sarasota Bay samples were extracted in two batches, one for faeces samples and the

other for gastric samples. Blank extracts from these batches were assayed with the

universal primers from chapter 2 and there was no indication of contamination in either

batch.

3.3.3 SWBD Universal Assays

A total of 470 clones were screened from 12 clone libraries with 29 – 52 clones screened

per clone library. At least one prey species was detected from all but one sample using

the universal method with an average of 3.2 +/- 1.7 (S.D) prey species identified per scat

(range 0-6; Table 3.3). All species of prey except for penaidae were recovered at least

once across all clone libraries (Table 3.3).

3.3.3.1. Clones of predator Origin in SWBD clone libraries

Chapter 4 investigates the NUMTs found in all clone libraries in terms of the sequence

diagnosis of NuMts and ramifications of these for DNA-based diet studies. In this chapter,

the data are presented to summarize the effect of predator origin clones on the efficacy

of this method for SWBD dolphin scats.

There were 10 putative NuMt variants isolated across all clone libraries with 8 appearing

in different samples and 7 in multiple individuals (Table 3.4).The proportion of predator

origin clones (i.e. T. truncatus mtDNA and nuclear mitochondrial pseudogenes – NuMts)

in lower volume samples was greater than in the relatively large volume samples used in

Chapter 2, suggesting sample volume may affect the efficacy of this method for these

dolphin scat samples (Fig. 3.4). The sparseness of this dataset, non-independence of

samples and the variation present in the 1.26 – 3.77 cm3 category meant these data

were unsuitable for modelling with logistic regression, however inference from 95%

confidence intervals of the mean in each volume category supports an effect of sample

volume on the efficacy of this method for these dolphin scat samples (Fig. 3.4).

Table 3.3 Raw data from universal prey detection assays of 12 SWBD samples. Samples with indicate this sample was assayed as part of the first chapter and the data are included here. Dashed lines indicate different days of the feeding trial.

Sample Volume (cm

3 )

Trial Day

Animal Time Sa Ag Mc N P Tn Sr Sl & Ss

Ps Tt Pg Total Clones

Screened

Prey Species

Detected

76 <1.26 5 Gemma 1130 0 0 0 0 0 0 0 0 0 0 29 29 0

77 1.26 – 3.77 5 Moki 1215 5 2 0 9 0 1 1 21 0 2 0 41 6

78 >3.77 5 Nila 1330 11 5 0 2 0 1 0 32 0 0 1 52 5

79 >3.77 5 Gemma 1400 0 1 0 0 0 4 0 16 0 4 9 35 3

81 >3.77 5 Gemma 1535 8 6 0 0 0 4* 1 17 0 1 0 37 5

91 >3.77 7 Gemma 1415 2 13 16 0 0 0 0 2 0 0 5 38 4

97 1.26 – 3.77 9 Stormy 815 1 0 14 0 0 0 0 0 0 0 24 39 2

99 1.26 – 3.77 9 Stormy 935 10 0 0 0 0 1 0 0 1 10 9 31 3

102 >3.77 9 Stormy 1355 5 0 11 0 0 12 0 0 4 6 1 39 4

104 Scraps 9 Buck 1410 0 0 1 0 0 0 0 2 0 6 25 34 2

112 Scraps 10 Stormy 945 2 1 1 0 0 0 0 0 0 19 25 48 3

118 1.26 – 3.77 10 Stormy 1135 1 0 0 0 0 0 0 0 0 17 29 47 1

Combined Totals

45 28 43 11 0 23 2 90 5 65 157 470

FOC* 9/12 6/12 5/12 2/12 0/12 5/12 1/12 6/12 2/12

Sa: Scomber australasicus, Ag: Arripis georgianus, Mc: Mugil cephalus, N: Nototodarus spp., P: Penaeidae spp., Tn: Trachurus novaezelandiae, Sr

Sillago robusta, Sl & Ss: Sardinella lemuru, Sardinops sagax, Ps: Pomatomus saltatrix, Tt: Tursiops truncatus, Pg: Tursiops truncatus pseudogene.* FOC

– frequency of Occurrence

51

Table 3.4 Nuclear mitochondrial pseudogenes (NuMts) variants isolated from SWBD clone libraries

SWBD NuMt Variant

Sample ID Individual 1 2 3 4 5 6 7 8 9 10 Unique Variants

76 Gemma Y Y Y Y 4

77 Moki 0

78 Nila Y 1

79 Gemma Y 1

81 Gemma 0

91* Gemma 1

97 Stormy Y Y Y Y Y 5

99 Stormy Y Y Y Y 4

102 Stormy Y 1

104 Buck Y Y Y Y Y Y Y 7

112 Stormy Y Y Y Y Y Y 6

118 Stormy Y Y Y Y Y Y Y Y Y 9

*Sequence was of poor quality although obviously originating from a Numt. The clone was not re-sequenced and subsequently not classified to any one NuMt variant.

Figure 3.4 Proportion of predator origin (mtDNA and NuMT) clones from different sample volume categories of SWBD samples. Sample volume categories and number assayed indicated on the x-axis. Filled circles are individual data points, the open circles the mean of each volume category, whiskers bootstrapped 95% binomial confidence interval for the mean for grouped data.

52

3.3.3.2 Results of SWBD Universal Assay Prey Detection compared to known diet

For all approaches comparing clone library results to known diet, predator derived

clones were excluded from consideration. Since the actual proportions of each of the

‘non-test’ prey species were unknown, these were pooled together in one category.

Simple frequency of occurrence (FOC) placed Scomber australasicus as the most

important prey item (9/12) followed by ‘non-test’ prey species (8/12), Arripis georgianus

(6/12), Mugil cephalus (5/12) and Nototodarus spp. (2/12) (Table 3.4). This is in contrast

to the actual importance by wet weight of the prey categories of ‘non-test’ species

(43%), equal S.australasicus, A. georgianus and M. cephalus (15% each) and Nototodarus

spp. (10%).

A weighted FOC score was calculated for each prey group where FOC was normalised to

the prevalence of clones of each prey species as a proportion of all prey clones across all

samples. By this scale ‘Non test’ prey were the most important prey group, followed by

S.australasicus, M. cephalus, A. georgianus and Nototodarus spp. (Table 3.5).

Table 3.5 Frequency of Occurrence (FOC) scores weighted by proportion of total prey clones for SWBD samples. Prey groups arranged in order from most to least important according to weighted FOC scores

Prey Group FOC Total Clones (%)

Weighted FOC Score

Ratio to 'Non-Test' Weighted FOC

Ratio to 'Non-test' by wet weight

‘Non-test' 8/12 48.6 0.324 - -

S.australasicus 9/12 18.2 0.137 1:2.4 1:2.86

M. cephalus 5/12 17.4 0.073 1:4.5 1:2.86

A. georgianus 6/12 11.3 0.057 1:5.7 1:2.86

Nototodarus sp. 2/12 4.5 0.007 1:43.6 1:4.3

Penaidae 0/12 0 0.000 - -

Mean prey proportions between all scat samples indicated the same relative importance

of prey items as the weighted frequency of occurrence approach (Fig. 3.5). Arithmetic

confidence intervals for mean prey proportions between scats incorporated the actual

percentage of prey in the diet for all of the fish prey, but not for Nototodarus spp. or

Penaidae (Fig.3.5). When considering averages across individual dolphins the relative

importance of prey items was Non-Test species, followed by M. cephalus, S.australasicus,

A. georgianus and Nototodarus spp. although the differences between M. cephalus and

S.australasicus averages were small (Fig. 3.5). Arithmetic confidence intervals for mean

prey proportions between individuals incorporated the actual percentage of the prey in

the diet for all prey items detected (i.e. all but Penaidae, Fig. 3.5). For between scat and

between individual averages, bootstrapped estimates of 95% confidence intervals of the

mean and of the 2.5 and 97.5 percentiles of 95% confidence intervals were very wide,

indicating that the process of sampling individual sample clone libraries adds

considerable variability to these estimates (Fig 3.5).

When all clone libraries were pooled and absolute prey proportions estimated, the

relative importance of prey matched the weighted FOC approach, although the

53

differences between S.australasicus, M. cephalus and A. georgianus were the smallest of

any of the above approaches (Fig. 3.5). Confidence intervals for pooled sample absolute

proportions incorporated the actual percentage of the prey in the diet for

S.australasicus, M. cephalus and A. georgianus but not for other prey groups (Fig.3.5).

Figure 3.5 Actual Proportion of prey in SWBD diet (Dark Grey horizontal lines) compared to proportions of amplicons in clone libraries calculated by different means. 1) Prey proportions averaged across individual scat samples (circles), 2) Prey proportions averaged across individual dolphins (squares), and 3) absolute total proportion (triangles) by pooling all prey amplicons in clone libraries. Whiskers represent the bootstrap 2.5 and 97.5 percentiles for 95% confidence intervals (light grey lines) and 95% confidence intervals for the mean (red lines) due to multinomial sampling error. Symbols represent arithmetic mean between groups (circles and squares) or total prey proportion (triangle). For comparison, arithmetic 95% confidence intervals of arithmetic mean proportions of raw data are given (dotted lines). Prey species as follows: Sa: Scomber australasicus, Ag: Arripis georgianus, Mc: Mugil cephalus, N: Nototodarus spp., P: Penaeidae spp

54

3.3.4 SWBD Mugil cephalus DNA assays

3.3.4.1 First and last detection of Mugil cephalus DNA

There were two independent positive detections of Mugil cephalus DNA 2.83 and 5.08

hours after the potential first ingestion of this prey ≈9am. All other scats produced on

the first day of the ‘Test’ period were of ambiguous origin.

The last independent detections of the prey signal were at 11:15am and 2:10pm the day

after the test diet ceased. Since the last possible feeding time for this prey was≈ 4pm

the day before, this suggests that prey signals can persist for a minimum of 19.25 –

22.16 hours after prey ingestion. Assuming the final signal was from prey consumed the

day before and given the first feeding time ≈9am these final prey detections could also

represent prey ingested up to 26.25 – 29.16 prior. Similar to minimum time to detection,

it is not possible to estimate the maximum prey signal persistence time, since positive

scats may have been produced, but not collected, after the last positive scat was

collected.

3.3.4.2. Positive samples from first to last prey detection

After the first prey detection between 33 – 80% of samples per day tested positive until

the last detection of the prey item (fig. 3.6). Of the 68 samples between the first and last

positive sample (inclusive) 60% were positive for the prey item. The bootstrap estimate

of the variance of the proportion positive was 6% (i.e. 60 +/- 11.9% (95% confidence

intervals); Fig. 3.7).

3.3.4.3. Effects of sex and sample volume on Mugil cephalus DNA detections

There were 42 samples collected between the first and last prey detection where the

animal producing the scat was certain. In an attempt to account for the serial

dependence of samples from the same individual, a mixed-effects logistic model was fit

to these data to examine the effects of sex and sample volume on Mugil cephalus assay

success (response: presence/absence of Mugil cephalus DNA) with individual as a

random effect. With an average of 5 +/- 3 (SD; range 1-10) samples per individual (n = 10)

data were too sparse to model (over dispersion). Ignoring serial dependence from

individual dolphins, pooling results from first to last prey detection in volume categories

resulted in data too sparse for robust contingency table tests, with expected cell counts

falling below 5 (Zar, 1999). Pooling data within sex, there was no significant difference in

Mugil cephalus DNA detection between sexes (X2=1.8, P = >0.1).

55

Figure 3.6 Samples that tested positive for Mugil cephalus DNA per day of the feeding trial (sequential days from left to right). For the first day, only samples collected after the first positive sample were included in this figure. Excluding the first test day, numbers above x-axis are total sample numbers for that day. Data are black dots with actual value to the right of the data point. Data are joined by lowess smoothed line.

Figure 3.7 Boxplot with overlaid probability density function estimated with kernel density estimates of proportion of positive samples from bootstrap replicates (n = 1000) from first to last prey detection. Filled red line is the actual proportion of positive samples over the duration of the prey detection period; dotted red lines are the bootstrap 95% confidence intervals of the proportion of positive samples calculated from the bootstrap variance estimate.

1 4 7 7 12 3 4 5 8 5 9 8 8 6 7 0 0

0.71

0.57

0.50

0.33

0.50

0.80 0.75

0.60 0.56

0.63

0 0 0 0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

NT NT Test Test Test Test Test Test Test Test Test NT NT NT NT

Pro

po

rtio

n p

osi

tive

Day of Trial from left to right

56

3.3.4.4. Evidence for prey pulse signals from Universal and Q-PCR assays

For samples assayed with the universal technique, there were three days where

sequential scats were collected from the same individual (Table 3.3). The first of these

consisted of three scats from a female (Gemma) where 0, 3 and 5 prey species were

detected from sequential scats 76, 79 and 81 respectively (Table 3.3).The second scat

was passed 2.5 hours after the first and the third 1.58 hours after the second with all of

the prey in the second scat detected in the third. The second day there were three scats

(97, 99, 102) from a male (Stormy) were passed with 2, 3 and 4 prey species detected

sequentially. The second scat was passed 1.3 hours after the first and the third 4.3 hours

after the second with one species from the first scat not detected in the second and all

species from the second detected in the third. The third day there were two scats (112,

118) from a male (Stormy) were passed with 3 and 1 prey species detected sequentially.

The second scat was passed 1.8 hours after the first and the single species detected in

the second scat was also detected in the first. In summary, there was substantial

variation in prey detection between sequential scats from the same individual indicating

prey DNA pulse signals: Both the appearance and disappearance of prey signals can

occur in sequential scats within 1.58 and 1.3 hours respectively.

For the Q-PCR results, there were 4 scenarios to consider the pulse signal of Mugil

cephalus DNA in sequential scats from the same individual: two no transition (i.e. no

detection -> no detection, or detection -> detection) and two transition (i.e. no

detection -> detection, or detection -> no detection). Three individuals showed no

transition passages (Gemma, four no detection -> no detection passages between 17m –

5.3 hours, Moki and Nila, one detection -> detection passage each, 1.58 and 3.7 hours

respectively) and one individual displayed a no detection -> detection transition passage

of 4.3 hours.

3.4 Sarasota Bay dolphin Universal Assays

A total of 485 clones were screened from 24 clone libraries, with 15 – 38 clones

screened per clone library. At least one prey species was detected from all samples

(Table 3.6), except one fecal sample which had a substantially lower volume than the

others (Table 3.2). The different bioinformatics routines used for prey identification in

this chapter yielded, in some cases, substantially different results compared to each

other and Chapter 2. For sequence quality control, in the case where no chimeras were

detected using CCode software, two chimeras were identified with ChimeraSlayer

software. As chimeras are not recognised with CCode software, there was an artificial

inflation of diet diversity estimates for the individual sample used in chapter 2. In both

samples, the chimeras consisted of different amplicon ends if two different prey species

identified from the sample, which may confound the chimera detection strategy used

CCode software. The sequences of predator origin identified appear to be NUMTs and

are the subject of Chapter 4, so they are not discussed in detail in this chapter.

57

3.4.1 Prey Assignment by BLAST threshold and Placement in Neighbour Joining

Analysis

There were 25 operational taxonomic units (OTUs) identified using the ‘percent

threshold/neighbour joining analysis’, with 20 OTUs assigned to species level

identifications and 13 of these displaying at least one OTU at a 100% match to GenBank

nucleotide records for that species (Table 3.6). There were an average of 3.2 +/- 1.7 (S.D.;

range 0-7) prey OTU’s identified per individual, with an average of 3.1 +/- 1.4 (S.D.;

range 0-7) OTU’s identified per scat sample and an average of 2.7 +/- 0.4 (S.D.; range 2-3)

per gastric sample (Table 3.6).

Table 3.6 Summary of OTUs identified from Sarasota Bay T. truncatus faeces and gastric samples using a Threshold/Topological Neighbour Joining Analysis

Prey OTU Identified (# of unique sequences)

FOC SBD

10* SBD 113

+

SBD 133

SBD 146

SBD 151

+

SBD 155

+

SBD 157

+

SBD 164

+

SBD 179

+

SBD 182^

SBD 195*

SBD 199

+

SBD 20

+

SBD 234*

SBD 236^

SBD 238

+

SBD 240

+

SBD 246*

SBD 90

Archosargus probatocephalus 1/19 10

Brevoortia patronus 2/19 1* 5 Centropomus undecimalis 1/19 7* Chaetodipterus faber 1/19 3 Cynoscion nebulosus 2/19 1 6

+*

Cynoscion sp. 1/19 2*

Elops saurus 4/19 4 4 2 11+

Epinephelus itajara 1/19 1

Gerreidae sp. 1/19 1 Hemiramphus brasiliensis 1/19 1

+

Lagodon rhomboids 12/19 18 31+* 25

+* 9 1 6 16 29

+* 2

+ 6 10 9

+*

Leiostomus xanthurus 9/19 10+* 1 1 4 4

+* 11 17

+* 3 3

+

Menticirrhus littoralis 1/19 7

Microgobius gulosus 1/19 2 Mugil cephalus 2/19 2

+ 1

Mugil gyrans/curema 3/19 2 2 2

Mugilidae spp** 1/19 1

Opisthonema oglinum 1/19 1 Opsanus beta 6/19 11 1

+ 2 9 5 9

+

Opsanus spp** 2/19 1 1+

Orthopristis chrysoptera 1/19 3 Paralichthys albigutta 1/19 3

+*

Prionotus sp. 2/19 1 2 Sardinella aurita 1/19 1 Sphoeroides parvus 4/19 1 6

+ 1 3

Predator Origin Clones 0 3 3 1 4 1 9 12 4 1 26 19 15 1 1 13 11 0 6

Total Clones Screened 20 19 40 41 19 20 20 20 20 37 38 19 20 20 39 19 19 20 35

OTU identifications 3 3 2 6 4 7 3 2 1 3 3 0 3 3 5 3 2 4 5

Indicates 100% match between at least one prey sequence and GenBank record; + Fecal Sample only; * Gastric Sample Only;

+* Both Fecal and Gastric

Sample; ^ Single clone variant in library chimera as identified with ChimeraSlayer.

59

3.4.2 Prey Assignment Using the MOTU/Statistical Assignment Package method

There were 53 unique sequences identified across all prey sequences which sorted to

28 taxonomic assignments consisting of 32 different MOTU’s using the statistical

assignment package (SAP): 13 MOTU’s were assigned to species due to 100% agreement

between the MOTU and database records and 5 MOTU’s were assigned to species level

by SAP (Table 3.7). Different MOTU’s were maintained on the basis of similarity to each

other allowing for Taq DNA polymerase error (See Table 3.7, Cynoscion sp., Scianidae,

Elops sp. and Opsanus sp.), and if below the ≥ 2 substitutions Taq error threshold, on the

basis of identical MOTUs in different samples (See Table 3.7, Scianidae 2 and Scianidae

3). There were an average of 3.7 +/- 2.2 (S.D.; range 0-9) prey MOTU’s identified per

individual, with an average of 3.3 +/- 1.8 (S.D.; range 0-9) MOTU’s identified per scat

sample and an average of 2.8 +/- 1.2 (S.D.; range 2-3) per gastric sample.

For some taxonomic groups there were few options for potential prey with distributions

that range into the Sarasota Bay area. For example, samples assigned to Elops sp. could

be only 1 of 2 species that are known from the Sarasota region; Elops saurus and Elops

smithi. The latter of this species is indistinguishable from the former without performing

counts of vertebrae. For the Lagodon rhomboides MOTU, some of the unique sequences

were assigned to Lagodon sp. in SAP, yet only Lagodon rhomboides occurs in the study

region. For the Opsanus sp. MOTU, although this sequence was substantially different

from Opsanus beta sequence records, Opsanus beta is the only Opsanus species known

to inhabit Gulf of Mexico waters. Considering the Scianidae assignments, there are 21

Scianidae species known to inhabit the Sarasota Bay region and 13 of these have

nucleotide records available on GenBank. It may therefore be the case that the

Scianidae MOTU’s represent one or more of the eight Scianidae sp. not represented on

GenBank. There is some difficultly in interpreting the Cynoscion sp. MOTU, since all

Cynoscion spp. known from the study area have multiple sequence records represented

on GenBank. It is noteworthy that for both the Opsanus sp and the Cynoscion sp.

MOTU’s which both have assignment interpretation difficulty, there are sequences of a

closely related MOTU, confidently identified to species within the same sample. It is

possible that these MOTUs are therefore experimental artefacts of the PCR process

outside of the normal bounds of Taq DNA polymerase error (e.g. prey mtDNA

hetroplasmy).

3.4.3 Comparison of different prey taxonomic assignment techniques

The difference in results between the two taxonomic assignment methods used on

recovered prey sequences originates largely from considering all sequences together

and accounting for Taq DNA polymerase error, as opposed to applying blanket cut-off

thresholds for each sequence when comparing it to database (i.e. GenBank) records. A

case in point is the different Scianidae MOTUs present in the SAP assignments, most

similar to Leiostomus xanthurus (Table 3.7). Using the threshold/neighbour joining

assignment method, all of these MOTUs grouped with Leiostomus xanthurus, since they

60

were all within 98% identity to this prey item (Table 3.6). By contrast using SAP and

accounting for Taq DNA polymerase error and multiple records from different samples,

these same sequences grouped into four MOTU’s which are probably more

representative of diet, given that 44% of the possible prey species in the family these

MOTUs could represent are not present in databases and intra-specific sampling in

databases is also low (i.e. < 3 separate records for all except Cynoscion spp.). Regardless

of whether Genus or familial level MOTU’s represent different prey species or intra-

specific variation, by representing them as MOTU’s they allow for the possibility they

could be different prey species or at least present a biological component of diet

diversity that can be compared between samples as a prey species might be. For this

reason further summaries of prey data from Sarasota Bay samples are presented with

the MOTU/SAP assignment method results.

61

...This page is intentionally blank...

Table 3.7 Summary of Sarasota Bay T. truncatus number of clones per individual taxonomically assigned to prey MOTUs using SAP software and accounting for Taq DNA polymerase error in clone libraries. If Genus and Family level identifications were > 95% similar to another identified taxon, they are listed below the most similar (% identity) species level identification and indented. Blue text indicated a difference in prey assignment between the ‘% Threshold/Neighbour joining’ method and the MOTU/SAP method.

SBD Number

Prey MOTU Taxonomic Assignment (Number of unique sequences if >1)

FOC 10* 20+ 90 113

+ 133 146 151

+ 155

+ 157

+ 164

+ 179

+ 182^ 195* 199

+ 234* 236^ 238

+ 240

+ 246*

Archosargus probatocephalus 1/19 10

Brevoortia patronus – (2) 2/19 1* 5

Centropomus undecimalis 1/19 7*

Chaetodipterus faber 1/19 3

Cynoscion nebulosus -- 2/19 1 6+*

Cynoscion sp. 1 1/19 1*

Scianidae 1 1/19 1*

Elops saurus (5) -- 4/19 1 3 4 1+

Elops sp. 1 1/19 1

Elops sp. 2 1/19 10+

Elops sp. 3 1/19 1

Epinephelus itajara 1/19 1

Eucinostomus sp. %

1/19 1

Haemulidae 1/19 3

Hemiramphus brasiliensis 1/19 1+

Lagodon rhomboides (9) -- 12/19 18 9+* 31

+* 25

+* 9 1 6 16 29

+* 2

+ 6 10

Leiostomus xanthurus (4) 6/19 3+ 7

+* 4 2

+ 12* 2

Scianidae 2 4/19 3* 1 11 1

Scianidae 3 (2) 3/19 1 2* 4+

Scianidae 4 1/19 1*

Menticirrhus littoralis 1/19 7

Microgobius sp. 1/19 2

Table 3.7 continued. SBD Number

Prey MOTU Taxonomic Assignment (Number of unique sequences if >1)

FOC 10* 20+ 90 113

+ 133 146 151

+ 155

+ 157

+ 164

+ 179

+ 182^ 195* 199

+ 234* 236^ 238

+ 240

+ 246*

Mugil sp. %

1/19 1

Mugil cephalus 2/19 2+ 1

Mugil gyrans/crema 3/19 2 2 2

Opisthonema sp. 1/19 1

Opsanus beta (4) -- 6/19 9+ 11 1

+ 2 9 5

Opsanus 1 %

2/19 1+ 1

Paralichthys albigutta 1/19 3+*

Prionotus sp. (2) 2/19 1 2

Sardinella aurita 1/19 1

Sphoeroides parvus 4/19 1 6+ 1 3

Total Prey MOTUs Identified 3 4 5 4 2 7 4 7 3 2 1 4 3 0 3 9 4 2 4

Total Predator 0 15 6 3 3 1 4 1 9 12 4 1 26 19 1 1 13 11 0

Total Clones screened 20 20 35 19 40 41 19 20 20 20 20 37 38 19 20 39 19 19 20

+ Fecal Sample only; * Gastric Sample Only;

+* Both Fecal and Gastric Sample;

^ Single clone variant in library chimera as identified with ChimeraSlayer

64

3.4.4 Occurrence and Inferred Relative Importance of Sarasota Bay dolphin prey

Simple frequency of occurrence (FOC) placed Lagodon rhomboides as the most

important prey item (12/19) followed by Opsanus beta and Leiostomus xanthurus (6/19),

Sphoeroides parvus, Scianidae 2 and Elops saurus/smithi (4/19), Scianidae 3 and Mugil

gyrans/crema (3/19), Prionotus sp., Mugil cephalus, Cynoscion nebulosus and Brevoortia

patronus (2/19) with all other prey items occurring in only one individual (Table 3.7). At

the familial level FOC placed Sparidae as the most important prey family (12/19)

followed by Scianidae (10/19), Batrachoididae (6/19), Mugilidae (5/19), Clupeidae,

Elopidae and Tetraodontidae (4/19) and Triglidae 2/19 with all other families occurring

in only one individual.

A weighted FOC score was calculated for each prey group where FOC was normalised to

the prevalence of clones of each prey species as a proportion of all prey clones across all

samples. By this scale all species level FOC assignments remained similar as basic FOC

rankings except for relative differences in each FOC ranking category (Table 3.8).

Weighed FOC of prey families remained similar to basic FOC, except Elopidae ranked

above Muligidae, and as with species, there were relative differences in each FOC

ranking category (Table 3.8).

Averaged prey amplicon proportions indicate that Lagodon rhomboides amplicons were

considerably more prevalent in clone libraries compared to any other species (Fig. 3.8).

Other prey species amplicon proportions show the same trend of relative prey

prevalence as weighted FOC scores, however there was considerable overlap in the 95%

confidence intervals of mean amplicon proportions for all prey species other than

Lagodon rhomboides (Fig. 3.8). Similarly, at the familial level, Sparidae amplicons were

considerably more prevalent in clone libraries compared to any other prey family.

Scianidae and Batrachoididae amplicons were the next most prevalent prey families at

an average of 20% and 11.2% of amplicons across individuals, respectively. All other

prey families were represented by on average <5.8% of prey amplicons with

considerable overlap in 95% confidence intervals.

3.4.5 Comparison Sarasota Bay dolphin diet estimated via DNA and stomach

contents data

Comparing data from this study to the six most frequenctly occurring prey taxa in Barros

& Wells (1998), FOC values were similar between the two studies with all but one

(Orthopristis crysoptera) within 10% of each other (Fig 3.9). Confidence intervals are

wide in both studies due to the low number of total samples analysed and show

considerable overlap in all species compared (Fig. 3.9). Averages of numerical

proportions of total prey in each stomach in the 12 most numerically dominant prey

from Berens et al. (2010) were also compared with data from Barros & Wells (1998) and

average total proportion of clones from these prey in this study (Fig. 3.10).

65

Table 3.8 Species and family Frequency of Occurrence (FOC) and FOC scores for taxa appearing in more than one sample. Taxa are grouped in descending order of weighted frequency of occurrence score.

Prey Group FOC Total Clones (%) Weighted FOC Score

Species

Lagodon rhomboides 12/19 43.3 0.273

Opsanus beta 6/19 9.9 0.031

Leiostomus xanthurus 6/19 8.0 0.025

Elops saurus/smithi 4/19 5.9 0.012

Scianidae 2 4/19 4.3 0.009

Sphoeroides parvus 4/19 2.9 0.006

Scianidae 3 3/19 1.9 0.003

Mugil gyrans/crema 3/19 1.6 0.003

Cynoscion nebulosus 2/19 1.9 0.002

Brevoortia patronus 2/19 1.6 0.002

Prionotus sp 2/19 0.8 0.001

Mugil cephalus 2/19 0.8 0.001

Family

Sparidae 12/19 48.5 0.306

Scianidae 10/19 19.7 0.104

Batrachoididae 6/19 11.0 0.035

Elopidae 4/19 5.9 0.012

Mugilidae 5/19 2.8 0.007

Tetraodontidae 4/19 3.1 0.007

Clupeidae 4/19 2.3 0.005

Triglidae 2/19 0.8 0.001

Of the 12 most numerically dominant prey from Berens et al. (2010), at least 9 and up to

11 taxa were also identified as prey in this study (range given since only inclusive higher

taxa were identified for some taxa in either study) and there were 8 common items with

Barros & Wells (1998) (Fig 3.10). Measures of variation for mean proportions were not

given in Berens et al. (2010), but were calculated from data provided in Barros and Wells

(1998) and in this study via the arithmetic mean of prey clone proportions across all

samples. In 4 cases the 95% confidence intervals for estimates of means from this study

encompassed mean estimates provided in Berens et al. (2010). In all instances

confidence intervals overlapped between Barros & Wells (1998) data and this study. The

mean diversity of prey items per individual with prey present was higher in this study

compared to Barros & Wells (1998): 3.9 +/- 2 (+/- S.D; range 1-9) and 2.4 +/- 1.4 (+/- S.D;

range 1-6), respectively, although estimates from this study could be overinflated if

MOTUs represent intra-specific variation. This information was not provided in Berens

et al. (2010). In terms of all taxa identified, Berens et al (2010) identified 22 species from

13 families, Barros & Wells (1998) identified 15 species from 10 families (and one

elasmobranch family) and this study identified at least 23 different species from 16

families. Of the species level identifications from this study, 5 species (Chaetodipterus

faber, Epinephelus itajara, Hemiramphus brasiliensis, Paralichthys albigutta and

Sphoeroides parvus) have not been found in the above prey identification studies for this

population.

66

Figure 3.8 Average taxon (top panel: species; bottom panel: family) amplicons per clone library. Filled circles represent arithmetic average across individuals, open circles represent proportion of pooled total prey amplicons. Whiskers are 95% confidence intervals for 1) population arithmetic mean (dotted lines of filled circles); 2) bootstrapped 95% confidence intervals for mean due to sampling error in each clone library (red line of filled circles) and 3) bootstrapped 95% confidence intervals for proportion of pooled total prey amplicons due to sampling error in each clone library (filled lines of open circles)

67

Figure 3.9 Absolute frequency of occurrence (sensu Wright, 2010) of six most common prey species in Barros & Wells (1998) compared to this study. Whiskers represent exact 95% binomial confidence intervals.

Figure 3.10 Average numerical proportion of prey items across all samples (other studies) and average proportion of prey item clones across all sample (this study). Whiskers represent truncated 95% confidence intervals of arithmetic mean. Variance was not available for Berens et al. (2010).

0

10

20

30

40

50

60

70

80

90

100

Lagodonrhomboides

Leiostomusxanthurus

Opsanus beta Mugil spp. Orthopristiscrysoptera

Elops spp

Ab

solu

te F

req

ue

ncy

of

Occ

ura

nce

(%

)

Barros & Wells (1998), n = 16

This Study, n=19

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Ave

rage

Nu

me

rica

l Pro

po

rtio

n o

f to

tal P

rey/

Clo

ne

s

Berens et al. (2010) n = 15

This Study

Barro & Wells (1998) n = 16

68

3.5 Discussion

3.5.1 Captive Feeding trials and SWBD prey detection results

Field based trophic ecology studies generally aim to determine population level diet

characteristics in the context of biologically relevant variation, such as spatial, seasonal

or ontogenetic factors. In such studies it is assumed that one or few scats are collected

from many individuals, in contrast to feeding trial experiments, where many scats are

collected from few individuals (Casper et al. 2007). This inevitably limits vertebrate

captive feeding trials, such as the feeding trial in this study. It is difficult to justify

treatment of each scat/DNA extraction from the same animal in feeding trials as

independent samples (e.g. Deagle et al. 2005; Deagle and Tollit 2006; Bowles et al. 2011;

OEHM et al. 2011), since previous studies have shown meal specific pulse signals for

particular prey items (Deagle et al. 2005; Casper et al. 2007). This implies time-series

correlation structure of scat derived DNA-diet data, which invalidates many traditional

statistical methodologies that require independence of samples (Zuur et al. 2009).

Although the nature of serial dependence is also of interest in terms of DNA gut passage

times and signal persistence, the constraints this places on inferences about results are

compounded by potential differences in digestive physiology between captive and real

world environments (Casper et al. 2007). These factors in combination with the under-

representation of inter-individual digestive variability (i.e. small n of different trial

subjects), regardless of potential environmental biases, and potential differences in pre-

ingestion DNA quality between live (i.e. predated) and dead (i.e. most items fed in

captive trial studies) prey, must temper the context of interpretation of feeding trial

data for DNA-based diet methods.

The SWBD feeding trial suffered from some of the constraints mentioned above, but

also had some unique limitations. Although the number of subjects was higher than

most feeding trials, the ability to identify them was poor: only ≈ 62% of samples

between first and last detection of Mugil cephalus DNA could be confidently assigned to

one individual. This resulted in small sample sizes which were not suitable for statistical

analysis. Other constraints were 1) that not all scats produced could be collected; 2)

judging by the fecal volumes collected from wild animals, only a small proportion of any

single scat was collected, and; 3) The times of ingestion of any particular prey item were

not recorded but were variable throughout the trial. Points 1 & 2 are in contrast to most

other marine mammal feeding trials, which were able to collect the majority of every

scat produced (e.g. Deagle et al. 2005; Casper et al. 2007; Bowles et al. 2011) and in

some cases, know the identity of the animals producing the scat (e.g. Deagle et al. 2005;

Casper et al. 2007). Indeed, the differences between the nature of the scats and the

time spent out of water between pinnipeds and cetaceans mean that comparisons

between this work and pinniped feeding trials may be misguided. The results of time

series scats from this study (3.3.4.1) also indicate prey (or meal) specific DNA signal

pulses. Taken together, the implication of these points for this study is that initial

detection and signal persistence cannot be ascertained from this experimental design. It

69

is also not possible to ascertain whether failure to detect prey items was due to lack

sensitivity of the molecular method (i.e. a false negative) or whether prey DNA was not

present to be detected (either in the entire bowel movement or in the portion of scat

collected). Treating the samples collected as a sample of all scats produced in the study

period, the detection rate of this study (60 +/- 12 %) was considerably lower than Deagle

et al (2005) and Parsons et al. (2005) and similar to the median value from three prey

species in Casper et al. (2005). Most samples collected in this feeding trial were small

volume, and sample volume appeared to be related to the proportion of prey amplicons

in clone libraries (Fig. 3.4). Deagle et al. (2005) found that prey DNA detection was less

consistent in small parts of individual pinniped scats as opposed to homogenized whole

scats and Oehm et al. (2011) also found that DNA extracted from relatively larger

volumes of avian scat had higher rates of prey detection than DNA extractions

performed on smaller homogenised samples. Another consideration is that these

samples were analysed upwards of 2.5 years after they were collected, which may not

be ideal for samples of marginal DNA quality to begin with.

With the exception of penaeid prey, the diet composition of SWBD samples was

accurately qualitatively defined using the universal technique. Penaeid DNA may not

have been detected due to a number of factors, including more severe DNA degradation

in dead prey pre-ingestion than other species, differential survival of prey species DNA

in the digestive tract, the fact that the item was not favoured by subjects and sometimes

found after feeding events in the water column (R. Casper pers. comm) or simply due to

variability in prey detection due to prey signal pulses coupled with the small overall

dietary proportion (2% wet weight) of the item. At the individual level it seems this

technique cannot inform as to all prey items ingested over the previous 1-2 day period.

However, at the population level where sample results are combined the technique

performed adequately for describing the major dietary components.

Like other studies, this work found a reasonable agreement between the weight of

proportions of prey ingested and the proportions of DNA signals from prey detected in

scats (Deagle et al. 2005; Deagle and Tollit 2006; Nejstgaard et al. 2008; Bowles et al.

2011). Where samples were analysed grouped by individuals or total pooled clone

libraries, the average estimated proportion of each prey group was within 11% of the

actual dietary proportion and in the case of grouping by individual dolphins, the 95%

confidence intervals for the arithmetic means overlapped prey wet weight proportions

for detected items. In comparison to grouping by scats, grouping data by individuals or

pooling is likely to be more representative of a field situation, where few scats are

collected from many individuals. Furthermore, individually grouped clone proportion

confidence intervals overlapped for items fed at similar proportions (10-15% wet

weight), but not for ‘non test’ items fed at 43% of wet weight, indicating that reliable

relative information can be gained by comparisons of prey metrics and confidence

intervals. The frequency of occurrence (FOC) and weighted FOC values did not compare

well in this regard and only appeared to be useful for comparison with amplicon

proportions estimates. The boot-strapping approach used to represent variance in

estimated prey proportions and their confidence intervals due to clone library sampling

error demonstrated the large variance introduced by sampling individual scat amplicon

70

pools. An inevitable constraint of cloning techniques is the time and expense of

sequencing individual clones and the consequences of this have long been recognised

for quantifying diversity in mixed (i.e. diet, environmental) samples (Hughes et al. 2001;

Dunbar et al. 2002). With the advent of next generation sequencing techniques such as

454 sequencing, there is no longer a requirement to clone samples to identify and

quantify amplicon diversity (e.g. Deagle et al. 2009). In addition to comparative ease of

this procedure, sample sizes of amplicon pools number in the thousand to hundreds of

thousands. As well as increasing the likelihood of detecting rare items, such large

samples of amplicon pools will also decrease variance in amplicon proportion estimates

due to sampling error. Despite this, it is still recommended that re-sampling or diversity

statistical summary procedures (e.g. Chao and Lee 1992) are implemented to

understand the impact of sampling error on study metrics and conclusions.

The feeding regime employed by this feeding trial is unlikely to represent any

experienced by wild animals, since all subjects were fed the same proportions of prey

each day. For extrapolation of results to field studies, it may be more appropriate for

captive studies to vary individual meals randomly each day and have the overall dietary

composition matching pre-defined proportions in feeding trials. In field studies the

individual scat is the sampling unit and it is the sample of these that is summarised per

group, rather than comparisons of individual scats themselves (Wright 2010). It may

therefore be more realistic to vary feeding trial diets per meal per individual, as this is

certainly more likely to be a realistic representation of the diet of wild animals. In any

case, it is also clear that cetacean feeding trials have their own unique set of practical

limitations, in that only small proportions of any scat can be collected and it is not

possible to collect all scats. Despite the caveats of this experimental design and the

inability to rule out false negatives, most prey species were detected. The ramification

of these results for interpreting data from wild animals are that although false negatives

cannot be ruled out, it is certain that prey species detected were consumed and, with

the universal technique used in this study, there is quantitative signal in the data - albeit

with wide confidence intervals. This is a significant advance for gathering trophic

information from these types of samples, as this was not previously possible at all.

3.5.2 Sarasota bay data

The application of the universal technique to Sarasota Bay samples yielded species level

prey information from all but one sample, but the bioinformatics routines implemented

to identify prey from DNA sequences yielded quite different results. An important issue

for using DNA sequence data to identify organisms is the effect of incomplete databases

on identifications (Nielsen and Matz 2006). While statistical procedures can be applied

to existing datasets to assign probabilities of inclusion to represented taxa (Nielsen and

Matz 2006; Munch et al. 2008b; Munch et al. 2008c), it is not possible to assign DNA

sequences to the correct taxa when they are not represented in databases (Munch et al.

2008a). Applying blanket similarity threshold cut-offs to existing data is appealing due to

the ease of implementation; however it is not representative of biological reality.

Implicit in a threshold value approach for taxonomic identity is that the cut-off value

71

incorporates both experimental error, such as Taq DNA polymerase error (in the case of

cloning), and intra-specific diversity. While the likelihood of experimental error fitting

into pre-defined dissimilarity thresholds may be appropriate, it is not likely these

adequately cope with variation in intra-specific diversity (Will and Rubinoff 2004).

Dissimilarity thresholds alone will therefore limit false positive results, but not

misidentifications (Ross et al. 2008), furthermore they cannot assign probabilities to

identifications. In essence a threshold approach was employed within the SAP

framework, since Taq DNA polymerase error was accounted for using an amplicon

length specific dissimilarity cut-off. Just as important to the final results were retaining

sequences that were identical between samples but which were under the threshold

cut-off value. These sequences likely represent common aspects of prey mtDNA

diversity. Prey mtDNA heteroplasmy may be responsible for these sequences,

particularly where a highly similar prey sequence was positively identified to species

level in the same sample. It is difficult to estimate the likelihood that this is the case

given the paucity of data on heteroplasmic mtDNA copy number variation in wild

populations. The SAP approach is appealing because it assigns a robust probability of

taxonomic inclusion based on Bayesian phylogenetic analysis. It therefore uses existing

data in a phylogenetic framework, instead of speculating on appropriate similarity

thresholds where there is inadequate sampling within and between species, as is likely

to be the case for most DNA-diet studies in non-model systems that use online

databases of reference sequences.

Failing bioinformatic assignment errors, there is the possibility of false positive results

from DNA sequences amplified from scat samples which merit consideration. False

positives may arise from, 1) environmental DNA contamination or cross-contamination

during the DNA extraction process, 2) Chimeric DNA formation in vitro, 3) mtDNA

heteroplasmy (as discussed above), 4) Nuclear mitochondrial pseudogenes (NUMTs -see

chapter 4) and 5) Where prey consumed by other prey and not the predator itself (i.e.

‘secondary predation’) are detected. Cross-contamination during DNA extraction,

chimeric DNA formation in vitro and NUMTs were all explicitly considered with

experimental controls and analysis methods. Environmental contamination may have

occurred for the scat samples, but seems unlikely for the gastric samples given that

these pass straight from a gastric collection tube to a sample vial. All but three scat

samples were collected aboard the research vessel, in most cases straight from the anus

of the individual and thus were unlikely to be contaminated. The three samples

collected in the water column yielded results qualitatively similar to vessel-collected

samples in terms of prey species detected, so it is not considered that they were greatly

affected by environmental contamination. Given the lack of detection of extraneous

DNA in the captive samples analysed (either environmental or secondary predation) it is

considered unlikely that this significantly affected results of the Sarasota Bay samples.

This position is congruent with the strong similarities between the DNA data and

previous stomach contents data from this population. However, given the different

circumstances of the captive and wild sample sets, false positives cannot be definitively

ruled out from the Sarasota bay data. Thus novel results presented, such as the

72

identification of new prey species or rare prey only occurring in small proportions in one

sample, should be interpreted with caution.

The adequacy with which the universal method described SWBD diet diversity does

suggest that the Sarasota data are qualitatively representative of actual Sarasota

dolphin diet. The diversity of diet items identified in this study is high compared with

past studies of dolphin diet based on hard-parts analysis (Barros and Wells 1998; Berens

McCabe et al. 2010). New dietary items were identified which may represent rare prey

or potentially those with hard parts unable to survive the digestive environment of

dolphin alimentary tracts for long. The SWBD data also suggest that averages of clone

library proportions should be broadly indicative of the relative mass of prey consumed

and sample diet proportions. These, in combination with FOC data, indicate that

Lagodon rhomboides (the pinfish) constitutes an important prey species for this

population. This is in strong agreement with a past study on the foraging ecology of this

population (Barros and Wells 1998).

In fact, there is a striking degree of similarity between data from this study and all past

works on the foraging ecology of this population. This is the first time a population scale

species level diet description from live cetaceans has been produced unconstrained by

the biases of other diet analysis methods, allowing comparison of diet gathered directly

from live free-ranging animals and stranded individuals from the same population. Not

only were frequencies of occurrence similar between this study and Barros and Wells

(1998), but mean proportions of prey clones and mean numerical abundances were

similar between this study and data from Berens McCabe et al. (2010). The DNA data

was intermediate between the two past studies for five of the 12 prey species where

these metrics were compared. There was also greater overlap between prey species

composition from Berens McCabe et al. (2010) and this study compared to both the past

stomach contents studies. Mean proportions of prey clones and mean numerical

abundances may not be directly comparable. However, considered together with prey

species composition they indicate that the DNA data are intermediate between the two

past stomach contents studies, suggesting that stomach contents diet data accurately

reflects diet composition and variation of live individuals in this population.

Bottlenose dolphins exhibit regionalized foraging behaviours in response to the

distribution, abundance and behaviour of their prey (e.g. Smokler et al. 1997; Connor et

al. 2000; Torres and Read 2009; Allen et al. 2011). It is hypothesized that Sarasota Bay

dolphins intercept sound from soniferous prey and use this to locate prey(the 'passive

listening' hypothesis; Barros 1993). The DNA data supports this hypothesis, both through

broad agreement of important prey taxa with the past studies, and the prevalence of

obligate soniferous taxa in four of the top 8 prey taxa from this study, despite their low

relative abundance in this ecosystem (data from Berens McCabe et al. 2010).

73

3.5.3 Conclusion

DNA-based diet analysis techniques were able to detect prey DNA in faecal samples

from live captive and free-ranging bottlenose dolphins. Using a specifically designed

universal prey detection assay, all known prey items except one (comprising only 2% of

the wet weight of the diet) were detected in 12 faecal samples from 5 different captive

individuals. The proportions of prey DNA were broadly congruent with the wet weight

proportion of prey items fed to the captive animals, however there was substantial

variation in prey DNA proportion estimates introduced by sampling error within the

molecular methodology. The small volume of most samples influenced the efficacy of

prey DNA detection using this method. The first detection of prey was 2.8 hours after

initial feeding and the signal persisted for a minimum of 19.3 hours after feeding ceased.

Between the first and last positive sample, approximately 60% of samples were positive

for the prey item. The detection rates from the feeding trial differed substantially from

past studies, highlighting some of the practical difficulties of performing robust feeding

trial experiments with cetaceans. Prey detected from samples obtained from 19 wild

free-ranging individuals consisted of up to 32 taxa across at least 13 families of teleosts.

The nature of the bioinformatics analysis applied to the recovered prey sequences

substantially influenced results and a Bayesian phylogenetic technique yielding

statistical confidence metrics for taxonomic inclusion of prey sequences was preferred.

The applicability of DNA-based diet techniques in situations where gaining specific diet

via other means information is impossible was demonstrated by the applying the

method to free-ranging dolphins.There was remarkable agreement between genetic

assessment of diet n free-ranging dolphins and past studies based on stomach contents

analysis of stranded animals in the same population. These techniques appear to hold

great promise for the analysis of cetacean diet, if samples can be obtained.

74

Chapter Four –

Pseudogenes and DNA-based diet analyses: a cautionary tale from a relatively well sampled predator-prey system

&

Chapter 5 –

Telomeres as age markers for marine mammals

These chapters have been removed for

copyright or proprietary reasons

Published as :

Dunshea, G., N. B. Barros, R. S. Wells, N. J. Gales, M. Hindell, A. and S. N. Jarman (2008) Pseudogenes and DNA-based diet analyses; a cautionary tale from a relatively well sampled predator-prey system. Bulletin of Entomological Research 98: 239-248. Garde E, A.K. Frie, G Dunshea , SH Hansen, KM Kovacs, C Lydersen (2010) Harp seal ageing techniques-teeth, aspartic acid racemization, and telomere sequence analysis. Journal of Mammalogy, 91: 1365–1374. Dunshea, G., D. Duffield, N. Gales, M. Hindell, R. S. Wells and S. N. Jarman (2011). Telomeres as age markers in vertebrate molecular ecology. Molecular Ecology Resources, 11: 225-235.

113

– General Discussion Chapter 6

Marine mammals, and particularly those without terrestrial phases in their life histories,

are challenging to study. Primary questions for both basic and applied ecology, such as

the role of marine mammals in the structuring of ecosystems and their dynamics, for the

most part remain contentious (e.g. Wade et al. 2007; Springer et al. 2008). Similarly,

habitat requirements, use, and key resources are ill defined for many marine mammal

species and particularly cetaceans. This all stems from the fact that these animals spend

the vast majority of their time, underwater and far away from land. In addition to the

scientific value of such information, this paucity of knowledge affects our ability to

understand the nature and scale of marine mammal interactions with anthropogenic

activities, and the impacts of these. This thesis has been concerned with developing and

evaluating molecular genetic techniques to indirectly examine specific aspects of marine

mammal ecology. The main aim was to find ways to maximise the information that can

be gained from samples, when the rare opportunity arises to sample these animals.

6.1 DNA-based methods to study cetacean diet

The notion that an animal's diet can be determined by examining its faeces is not at all

new (e.g. Storr 1961), but the methods to do so have changed. Although procedurally

complex, the conceptual simplicity of the DNA-based diet method is that it requires only

DNA to partially survive the digestive process. This is also one of the advantages of this

technique compared to examining faeces visually, where recognisable diet remains are

required, as is the expertise to identify them. As opposed to the specialized skills,

experience and reference collections required for visual faecal analysis, workers with

molecular biological skills and appropriate facilities are practically a ubiquitous feature

of contemporary research institutions (Bowles et al. 2011). For many different animals,

some or all dietary items are digested beyond any visual recognition (Symondson 2002b).

Cetacean faeces fall into this category, and without genetic techniques contain little

dietary information. Other uses for cetacean faeces includes individual hormone profiles

(Rolland et al. 2005), parasitic prevalence (Aznar et al. 2011) and as a source of DNA for

the animal itself (Parsons et al. 1999; Green et al. 2007) . Although DNA-based diet

methods had been applied to cetacean faeces previously (Jarman et al. 2002; Jarman et

al. 2004), this is the first study to examine their efficacy systematically and apply them

to a wild population to address an outstanding question for cetacean stomach contents

analysis.

The results from the captive trials highlighted some of the technical as well as practical

difficulties of captive feeding trials for determining cetacean diet. Many of these

difficulties constrained interpretation of the feeding trial data, but they paradoxically

may also represent the ‘best case’ scenario for the majority of studies sampling wild

cetacean scats. It is rare that cetaceans are captured at all, and rarer still that faeces are

collected during this, as with the Sarasota Bay research program. It is more likely that

114

use of these techniques for studies of wild cetaceans will generally require sampling

from the water column. Sampling from the water column needs careful consideration of

appropriate control samples to test for the possibility of microscopic organism

contamination (King et al. 2008). While the aquarium water is filtered and UV sterilized,

making it unlikely that fish larval or environmental DNA contamination occurred, it is

possible that contamination from cellular debris from prey fed would contaminate

samples.

Although the issue of microscopic organism contamination has been raised previously

(King et al. 2008), it is not clear what constitutes appropriate control samples. Simply

collecting a water sample with the faecal sample has inherent assumptions about the

distribution of zooplankton, and it is not clear whether the presence or absence of an

item in a control sample should preclude the presence or absence of the same item in

the faecal sample. For example, how many water samples should be collected and in

what vicinity to the faecal sample? It is known that zooplankton distribution is random

and varies over small temporal and spatial scales (e.g. Jenkins 1988) and therefore it is

currently not clear what form appropriate control samples should take. Additionally, if

the predator under study feeds on zooplankton it is likely that their prey would be

present in the water column where samples are collected. Therefore control samples

would not be useful for differentiating prey from environmental contamination. It is

assumed in this study that the sequences and templates detected in faecal samples

were from ingested prey with further discussion of these results.

The captive feeding trial demonstrated that prey DNA fragments up to 270 bp can be

detected in dolphin faecal samples. The composition of these fragments within

individual samples did not represent the breadth of diet over the previous 1 – 2 day

period; however across all samples 9 of 10 prey items were detected. That

undetermined, but small, fractions of individual bowel movements were collected

probably strongly influenced prey detection rates within samples, given the pulse nature

of the DNA signal from individual meal items shown in this and other studies. Sample

volume did affect the efficacy of the universal method for detecting any prey DNA, it

therefore follows that it affected detecting all prey DNA, in addition to the pulse nature

of prey DNA signals. The proportions of prey DNA in samples estimated with the

universal technique and summarized across samples reflected proportions of some prey

items (particularly fish) ingested, albeit with large error due to the hierarchical nature of

the sampling processes within the analysis of samples. That is, sampling of bowel

movements during the day, sampling within bowel movements (i.e. material actually

collected) and sampling of amplicon pools produced from PCRs via sampling clone

libraries. Quantitative relationships of prey DNA proportions reflecting the mass of prey

ingested have been reasonably inconsistent across different controlled feeding studies,

but different molecular methods and summaries of data across samples have been

applied to each (Deagle et al. 2005b; Deagle and Tollit 2006; Nejstgaard et al. 2008;

Bowles et al. 2011). Averaging or pooling across samples likely reduces the effects of

prey DNA pulses in individual samples and individual predator heterogeneity in digestive

processes, in addition to decreasing chances of a type II statistical errors. Squid

proportion in the diet was consistently underestimated by any data summary method in

115

this study, which could be related to differences in mtDNA density/copy number

between fish and cephalopods. It is likely that more distantly related taxa are more

likely to have increasingly different mtDNA densities for any given mass of tissue, which

could explain the similarity of fish DNA proportions, as opposed to squid, to each other

and the mass ingested. This could be accounted for by developing correction factors for

tissue mass specific copy number differences between prey, as has been attempted

elsewhere (Bowles et al. 2011). For future studies on wild animals where correction

factors are not available, interpretation of prey DNA proportions seems problematic

without some a priori information, both on the molecular technique being employed

and on prey inter-specific copy number and detection rate variation. As a starting point,

comparisons may be best reserved by examining proportions within constrained

phylogenetic groups, although even this may be problematic. In their feeding trial,

Bowles et al. (2011) found that one fish in mixed diets of four fish species was

consistently underestimated despite application of correction factors. A reasonable

approach may be that workers should use prey DNA proportions as ‘working hypotheses’

and treat them with wide confidence intervals, until further information can be obtained

by molecular experimentation or other diet analysis techniques.

It may be reasonable to assume that the individual level data may not represent all prey

items consumed within the previous 1-2 days, but the summaries across all samples

adequately reflect the vast majority of dietary items consumed by weight in these

individuals. Although beyond the scope of this study, some simple modelling could

examine this under a variety of scenarios. For example, simulations of feeding trial data

can inform how many samples are required to detect specific prey in a multi-species diet

with certain probabilities of false negatives (e.g. Casper et al. 2007). This framework

could also be extended to determine minimum sample sizes required to detect, with a

given statistical power, diet differences between populations (see Trites and Joy 2005).

Since far greater volumes of faecal material were collected from Sarasota Bay animals, it

may be that individual results are more indicative of all prey consumed than for the

captive feeding trial samples. Regardless, these results are significant for two reasons: 1)

Excluding observational reports, they are the first description of the specific dietary

composition of a population of live free-ranging cetaceans, and, 2) They are the first

data to allow direct comparison of diet from stomach contents of dead individuals with

the diet of live free-ranging individual cetaceans from the same population. This

comparison has therefore provided a separate validation of the conclusions of past

studies regarding the diet and prey selection preferences of this population, and in

doing so has provided further support for the proposed mode of foraging hypothesized

for this population (i.e. the ‘passive listening hypothesis’).

While the universal diet technique developed and employed throughout this thesis was

powerful for detecting prey DNA for known and unknown situations, it had

complications. The first was that the technique could not exclude unknown predator

nuclear mitochondrial pseudogenes (numts). This was more of an issue for samples

from the feeding trial than for Sarasota Bay samples: numt’s were sequenced in 10 out

of 12 faecal samples from the captive feeding trial and 13 of the 15 Sarasota Bay faecal

116

samples. The origin of these sequences needed careful consideration in light of potential

prey and mammalian 16S mtDNA sequence diversity to be properly assigned as numts.

For cetaceans this type of analysis is likely to be more straightforward then for lesser

known predatory taxa due to the availability of extant cetacean mtDNA sequence data

and mammalian mtDNA sequence data. But the ramifications of numts on DNA-based

diet analyses will depend on the strategy employed to detect prey DNA. Where DNA

sequences are the end results of prey detection methods, spurious matches to database

sequences should be considered with the utmost caution and should be examined for

potential numt origin as a matter of course. Where prey DNA sequences are not the end

result of prey detection methods, predator or prey numt amplification and scoring can

confound these results and this requires careful consideration in the design and

validation of these methods.

Despite repeated pleas over the years for laboratory and analytical strategies that can

detect or avoid numts when using mtDNA markers in ecology and evolutionary biology

(Sorenson and Quinn 1998; Bensasson et al. 2001b; Thalmann et al. 2004b; Thalmann et

al. 2005; Triant and Dewoody 2007; Song et al. 2008), many studies employing these

markers are still being published with no mention of these (Morin et al. 2010). This is

also the case in the DNA-based diet literature (e.g. Bohmann et al. 2010) despite

recognition of the ramifications of errors in mtDNA-based taxonomic identifications as a

result of mtDNA heteroplasmy or numts (Rubinoff et al. 2006). Clearly this is an issue

that working molecular ecologists, publication authors, journal editors and article

reviewers should all be aware of to rigorously perform their roles, although it is

currently not clear that this is the case.

This work has provided guidelines and proof of the efficacy of these nascent trophic

ecology methods for studying cetacean diet and foraging ecology. Other methods

applicable to live cetaceans have particular advantages and disadvantages, dependent

on the timescale of interest and prey taxonomic resolution required. Some of the

characteristics, relative advantages and disadvantages of other diet analysis techniques

for live animals are given in Table 6.1. Stable isotope (SI) analysis can yield a variety of

ecological information on live animals such as diet and trophic level information, habitat

use and migratory patterns, pathways of food web contaminants and physiological

processes (Newsome et al. 2010). Generally, only skin and blubber can be practically

sampled from most live cetaceans, however some exceptions exist (Blood or teeth;

Wells et al. 2004; Barros et al. 2010 respectively). Although low in taxonomic resolution,

SI techniques offer a temporal window of days to years for samples from live animals,

dependent on the tissue used (Table 6.1). This wide range in timescales has obvious

advantages for long/large scale questions such as ontogentic dietary shifts, variation in

foraging strategies and the effects of inter-annual variation in foraging success on

growth and breeding. The complementary ecological information gained from SI analysis

is also appealing as it is often directly related to the foraging ecology of the focal species.

A recent derivative of SI analysis called compound-specific stable isotope analysis

characterises stable isotopes at the level of individual organic molecules (Newsome et al.

2010). This new technique has the potential to indicate dietary sources at the base of

food webs, allowing disentanglement of spatial and trophic/physiological processes in

117

isotopic signatures; an important issue for migratory marine mammals (Newsome et al.

2010).

Fatty acids signature analysis (FASA) and its derivatives (‘quantitative fatty acids

signature analysis’ or QFASA; Iverson et al. 2004) characterise the fatty acids present in

the lipids of animals. The basis for FASA is that for mono-gastric predators, fatty acids

consumed by predators are deposited without modification, allowing inference of their

dietary intake over time (Budge et al. 2006). However, this very basis is not without

controversy, with some authors believing that deposition of prey fatty acids in predator

tissue occurs with modification and is unpredictable (Grahl-Nielsen et al. 2004). Prior

information on modification and synthesis of particular FA can alleviate this problem,

but is difficult to obtain in many instances (Grahl-Nielsen et al. 2004). In principle,

QFASA provides quantitative estimates of specific prey items consumed, however this

requires extensive a priori characterisation of variation in prey fatty acids and predator

metabolism (Budge et al. 2006). This has rarely been performed as it requires extensive

experimentation, development and validation of complex statistical models (Iverson et

al. 2004). More typically, FASA has been used to qualitatively describe intra-specific

variation in diet (e.g. Bradshaw et al. 2003; Newland et al. 2009), prey-base

separation/overlap in sympatric predators (e.g. Iverson et al. 2007) and spatial and/or

temporal patterns in diet (e.g. Wang et al. 2009). The timescale over which FASA

represents diet, days to months, is dependent on the tissue analysed and the ecology of

the predator, although few estimates of this are available for specific predators (Budge

et al. 2006). Like stable isotopes, these techniques therefore yield longer term dietary

data (Table 6.1), however data interpretation can be constrained by a lack of prior

information on prey fatty acid variation, predator metabolism and the appropriate

timescale for consideration.

In comparison, DNA-based dietary analysis facilitates high resolution taxonomic

identification of prey, but over comparatively very short timescales (Table 6.1). Thus,

aside from technique specific constraints, the trade-off when considering the best

technique revolves around the timescale of the study question and the taxonomic

resolution of prey required. Clearly, if questions are centred around specific trophic

interactions then specific techniques (i.e. DNA or in some instances FASA) are required.

However, if questions centre on long term foraging strategies and their effects on other

variables, a longer time-frame, but less taxonomically informative indicator of trophic

processes may be more suitable. For questions of resource partitioning, competition or

spatial and temporal segregation of prey sources within and among predatory taxa, each

technique can be informative over the appropriate temporal window. Coupling DNA

techniques with those with a wider temporal range is potentially a powerful

methodological system to understand the dynamics of the specific diet of predators. If

the timescales of integration of dietary fatty acids and/or stable isotopes can be

estimated, and longitudinal DNA-diet samples can be collected, there is potential to

match the lag in dietary integration of prey biomarkers with specific prey data collected

from around the lag time.

118

Of primary importance in these methodological considerations is the life history of the

study species. For example, most mysticete cetaceans undertake extensive seasonal

migrations from high latitude feeding grounds segregated from low latitude

breeding/calving grounds (Clapham et al. 1999). It is generally assumed that feeding

during migration and in breeding/calving grounds is minimal, especially for lactating

mothers, although there are few data to verify this (Clapham et al. 1999; Oftedal 2000).

At any rate, it is assumed that the most important nutritional input for these animals

occurs during the time spent on the high latitude feeding grounds (e.g. Leaper et al.

2006). Longer time-frame diet analysis techniques are thus more appropriate for

samples collected from the breeding/calving grounds of these taxa. For other taxa that

have less pronounced habitat segregation, shorter range dietary techniques such as

DNA-based methods are as relevant as longer term methods. Given that sampling of

mysticete cetaceans on feeding grounds is rare and DNA-based diet methods would be

prone to environmental contamination of potential food sources, it seems more likely

that DNA-based diet techniques will find more use in studies of odontocete cetaceans

and other marine mammal species.

Table 6.1 Summary of methods available to study the diet of live cetaceans

Method Timescale Prey Taxonomic Resolution (Qualitative/Quantitative diet signatures?)

Advantages Disadvantages

Stable Isotopes Variable dependent on tissue used: Blood Plasma – Hours to Days

[1]

Epidermis – One to 3-4 months[2]

Muscle or whole blood – A Few to Several Months

[1] Bone Collagen or teeth –

Several Months to Years[1]

Baleen - Years

[1,3]

Very Low – Can only indicate relative trophic position of prey (This may change with compound-specific analyses) (N/Y)

- Variable, but relatively long timescales for inferring processes through time - Unlikely to be confounded by secondary predation

- Low taxonomic resolution - Potential confounding of diet/trophic processes with movement or age/physiological/growth processes - Isotopic enrichment needs characterisation for robust interpretation of data

Fatty Acid Analysis

Variable dependent on species ecology and tissue used

[4]

Hours- Months[4]

Intermediate – Can sometimes discriminate between prey higher taxa (Y/N)

- Variable, but relatively long timescales for inferring processes through time - Large multivariate datasets powerful for ordination techniques

- Timescales of dietary integration are not well characterised for most tissues in most species

[4]

- Sampling process can be inherently subjective

[5]

- Complex data sets and interpretation - Potentially confounded by secondary predation

[4]

Quantitative Fatty Acid Signature Analysis

Variable dependent on species ecology and tissue used

[4]

Hours - Months[4]

Very High – Species Level (Y/Y– in some instances where prior information is available)

- Variable, but relatively long timescales for inferring processes through time - Large multivariate datasets powerful for ordination techniques

- Timescales of dietary integration are not well characterised for most tissues in most species

[4]

- Substantial prior knowledge is required through experimentation and characterisation of prey variability

[4, 5]

- Potentially confounded by secondary predation

[4]

Observation of Feeding

Present moment Variable – Can be species, usually at least Phyla level (Y/N)

- Insights into hunting and prey capture strategies can be gained

- Inevitably biased towards surface events

DNA-based Analysis

Previous 3-6 to 19 - ≈48 hours Very High – Species level (Y/potentially – in some instances where prior information is available)

- High taxonomic resolution - Targeted sampling of potential unknown prey (phylogenetic information in recovered DNA sequences)

- Database of prey DNA sequences is required for specific identification - Very short timescale of information - Potentially confounded by secondary predation or environmental contamination

[1] Rubenstein

and Hobson (2004);

[2] Based on estimates of odontocete epidermal turnover cited in Ruiz-Cooley et al. (2004);

[3] Best and Schell (1996);

[4]

Budge et al. (2006); [5]

Grahl-Nielsen et al. (2004)

120

6.2 Live marine mammal age estimation and telomeres

It would appear that age estimation for live marine mammals currently remains elusive,

however some recent work has provided potential new directions. Herman et al. (2008;

2009) have demonstrated strong predictive relationships between blubber fatty acid

metrics and age in free ranging killer whales and humpback whales. Although these

relationships are empirical and the underlying biological processes uncertain, their

accuracy and precision is relatively high. They also appeared to be applicable

independent of diet, sex or ecotype in killer whales but were probably less general in

application to humpback whales (Herman et al. 2008; Herman et al. 2009). There have

also been promising examples of using transcriptional profiles to estimate the age of

mosquitos (Cook et al. 2006; Cook et al. 2007). Despite these promising initial results,

until an understanding of the mechanistic biological processes responsible for these

results occurs, it remains difficult to validate their applicability for unknown individuals

in field situations. Many of the issues raised in this discussion regarding the influence,

interactions of, and variation in endogenous and exogenous variables on telomere-age

relationships are applicable to any biomarker approach.

This study highlighted some difficulties in telomere measurement methodologies and

conceptual problems in the notion of using telomeres for age estimation. It is clear that

there are issues in applying denaturing terminal restriction fragment (TRF) telomere

measurement methodologies to non-model species. I recommend that, genome-wide

telomere sequence should be characterised with Bal 31 time series experiments, or

other telomere sequence characterisation procedures such as cytogenetic procedures.

Without this, it is not clear that what is being detected and measured in denaturing TRF

assays is actually telomere sequence. If not, there is potential for whatever genomic

element that is being measured to co-vary with the processes under study (i.e. sexual,

age or individual or population specific differences) and provide non-random results that

have nothing to do with telomere length variation and telomere dynamics processes.

There are high profile telomere studies reporting novel evolutionary processes, where

there are clear (and potentially non-telomeric) banding patterns throughout the size

range of the ‘telomere’ fragments measured (Olsson et al. 2010; Olsson et al. 2011).

In light of Chapter 4 and the many other recent telomere studies, a reappraisal and

critical review of the notion of using telomeres for age estimation was undertaken. Here

the studies advocating age estimation via telomeres were assessed against contradictory

work. This involved considering processes influencing telomere dynamics as well as

experimental design issues. It was concluded that despite earlier work, the weight of

evidence now indicates that telomeres alone should not be used for age estimation

without considerable further work.

6.2.1 Telomeres for age estimation: Reappraisal of the literature

The best examples of where telomere length (specifically TRF measurements) may be

used to estimate age of wild animals come from birds. Birds are ideal candidates for

telomere length analysis: samples of blood can be relatively easily obtained, cells can be

121

effectively buffered immediately to prevent DNA degradation and as little as 50–500µL

of blood can yield upwards of 10µg of DNA due to avian red blood cells being nucleated.

Thus, there is optimal collection, storage and DNA yield conditions for samples of bird

blood. These studies (Haussmann and Vleck 2002; Haussmann et al. 2003b) were able to

differentiate age into three lifespan stages in all measured individuals in all species

studied except zebra finches (Vleck et al. 2003), although a few statistical details are

given on this.

Attempts to improve the accuracy of age estimates gained from TRF analysis

(Haussmann and Mauck 2008a) involve selectively analysing size classes of telomeric

DNA that increase the correlation coefficients of relationships between telomere length

and known age samples. These methods were able to improve correlation coefficients (r 2 values) from <0.2 to >0.8. Yet, at the best correlation that could be found between

telomere length and age, there are data points where telomere metrics overlap

between 6-month and 2.5-year-old individuals. Zebra finches have a maximum lifespan

of more than 5 years (Haussmann et al. 2003b); thus, for individuals, this age resolution

is unable to distinguish an individual approximately one-tenth through its life from one

halfway through its life. This level of predictive accuracy for the ‘best’ possible situation

for individual telomere-based age estimation has obvious consequences for the

downstream use of the data. This also raises the issue of the use of the telomeres on a

‘population age structure’ scale rather than an ‘individual age estimate’ scale, which is

addressed later.

The logic employed by Haussmann & Mauck (2008) is also questionable. They used

samples of known age birds and numerically determined the size classes of telomeres to

analyse that yield the tightest relationship between age and the telomere metric for that

sample. As the premise of the relationships is to use them to predict age in unknown

individuals, this tactic assumes that the estimates of the best size class of telomeres to

analyse are independent of the characteristics of the individuals used in the analysis. It is

not clear that this is the case and it cannot be assessed from the experimental design. If

this assumption is not satisfied, the predictive relationships should only be used on

animals from that particular sample set, which are already known age. The authors

provide an explanation for why focusing the telomere size range selected numerically in

their analyses should be appropriate for unknown samples; specifically, that accruing of

short telomeres is a marker of older age. However, if these telomere size classes are

most illustrative of age and this is constant (an assumption when applying these

relationships to unknown individuals), there should be no need to calibrate this on a

species, and indeed sample-specific basis.

Another recent study has found that telomere data are useful for age estimation of

martens (Martes spp.; Pauli et al. 2011). This study is a tremendous conceptual advance

to this field because it acknowledges the effect of influential exogenous covariates on

telomere dynamics and attempts to include these in a Bayesian network statistical

analysis. Despite this, the study suffers from a lack of reporting of technical details of

qPCR assays, however it is not alone in this regard (see Horn et al. 2010). There are also

aspects of the data analysis that are unclear and problems with data interpretation. The

122

authors used muscle blocks from tooth-aged, harvested animals as a tissue source to

generate telomere data, developing models for age class estimation with covariate

datasets available during ‘non-invasive’ sampling (i.e. collection of skin or plucked hair

samples) or ‘live caught’ sampling, presumably where non-destructive sampling is

desirable. Unfortunately telomere length and dynamics are not homogeneous among

tissues (Friedrich et al. 2000; Nakamura et al. 2002), therefore it is unlikely that the

application of their models to telomere length from non-muscle tissue is appropriate.

These authors also argue that interstitial telomere sequence is unlikely to affect their

results because it is correlated with terminal restriction fragment data in other species.

As stated previously, this will be dependent on the genomic characteristics of ITS, which

can vary over small phylogenetic distances (Meyne et al. 1990). Additionally, Pauli et al.

(2011) used covariates that were likely to be strongly influential on the age structure of

their sample as priors in the Bayesian modelling framework, rather than those likely to

affect telomere dynamics. Given that there are no mechanistic links between their

covariates with telomere dynamics processes, it is not clear whether the nature of the

relationships between variables used for model parameterisation are representative of

the effects of more important variables for telomere dynamics of these species. These

authors also used the qPCR technique and generated relative telomere metrics in

relation to an arbitrarily chosen animal. Therefore, all future application of these models

rely on comparisons with this same arbitrarily chosen animal, since the telomere metrics

used to parameterise the predictive model were all created with reference to this

individual. Lastly, although sensitivity analysis was performed on model outputs, this did

not incorporate uncertainty in telomere metrics determination, which was a

(presumably, mean) co-efficient of variation of 14.6% from a single run of duplicates of

each loci for sample. There was no consideration in variation of metrics among runs, as

was noted in this study, in Pauli et al. (2011). These drawbacks highlight some of the

complexities involved in the notion of using telomeres as age markers and the need for

interdisciplinary approaches. However, despite these drawbacks, this style of study and

data handling will likely be most useful for telomere based age estimation studies. This

will of course be limited to study taxa where information on appropriate covariates will

be available, which is unlikely for many marine mammal taxa.

Other studies have concluded that telomere metrics are not useful for age estimation

(Hall et al. 2004; Juola et al. 2006; Karlsson et al. 2008; Garde et al. 2010) and found no

relationship between telomere length and age (Hatase et al. 2008; Bize et al. 2009).

These, and other studies, have shown large variation in telomere length within age

classes and telomere dynamics between individuals at similar developmental stages and

ages in a wide range of vertebrate taxa and using a wide range of methods (Hall et al.

2004; Juola et al. 2006; Pauliny et al. 2006; Aviv et al. 2008; Hatase et al. 2008;

Haussmann and Mauck 2008b; Bize et al. 2009; Nordfjall et al. 2009; Salomons et al.

2009), highlighting the importance of both genetic and environmental factors in

determining initial telomere length and rate of attrition. This betrays the context-

specific nature of relationships between telomere length and age. For example,

favourable relationships between human telomere length and age were reported for

forensic age estimation (Tsuji et al. 2002). However, similar relationships derived with

123

similar methods but different study subjects show very different models for human

telomere–age relationships (Unryn et al. 2005; Valdes et al. 2005). This suggests that the

relationships are, at best, useable for the populations they were developed from, but

not for different populations. While it was initially thought telomeres may provide an

age marker for forensic use (Tsuji et al. 2002; Takasaki et al. 2003; Nakagawa et al. 2004),

this has been rejected because of issues of reliability and error of age estimates

(Schmeling et al. 2007; Karlsson et al. 2008). These data suggest that, not only are there

influences other than chronology on telomere dynamics, but that these also vary

between individuals and subject groups. For wild animals, population- and cohort-level

variations are also expected.

There a number of endogenous and exogenous factors that affect telomere dynamics

(see Monaghan and Haussmann 2006; Aviv 2008). The level of telomerase activity in

tissues obviously has a profound and tissue-specific influence. Here, I focus on the

factors that vary at the individual level, which may influence initial telomere length,

telomere attrition rate, telomerase activity or all of these. Initial telomere length, while

being greatly variable, has a heritable component (Slagboom et al. 1994; Rufer et al.

1999; Nordfjall et al. 2005; Andrew et al. 2006) and in human offspring is positively

linked to paternal age (Unryn et al. 2005), although the extent and nature of heritability

is uncertain (e.g. Bischoff et al. 2005). There also appears to be heritable aspects to

regulation of telomere dynamics in vivo (Graakjaer et al. 2004). Telomere attrition rate is

dependent on average telomere length, with individuals with relatively long average

telomere length displaying more telomere attrition over time (Aviv et al. 2008; Bize et al.

2009; Nordfjall et al. 2009), although care is needed interpreting these data to avoid

regression to the mean artifacts (Kelly and Price 2005). It also appears that this relatively

higher rate of attrition acts preferentially on the longer telomere size classes (Salomons

et al. 2009). This is intriguing considering that longer telomeres are generally considered

to be a proxy for better individual quality or condition (e.g. Pauliny et al. 2006); it

suggests that the presence of longer telomeres at old ages is indicative of relatively

increased telomere maintenance mechanisms as much, or more than, decreased loss

through processes that effect attrition. The above longitudinal studies and others (Hall

et al. 2004; Ujvari and Madsen 2009; Chapter 5) also demonstrate that individuals can

exhibit stable or increased average telomere lengths over considerable periods of their

lifespan. Apparent telomere lengthening has appeared in a number of different

longitudinal studies on a number of different taxa with a variety of methods, suggesting

it may be a real phenomenon and not measurement error (see Salomons et al. 2009).

Telomere lengthening has considerable importance in relation to interpretation of cross-

sectional data and the models fit to it.

There are also exogenous factors that influence telomere dynamics, albeit through their

effects on endogenous processes. Telomerase activity and telomere dynamics have

been linked to stress levels in humans and mice, with greater stress associated with

shorter telomeres and decreased telomerase activity in otherwise healthy women (Epel

et al. 2004) and shorter telomeres in mice (Kotrschal et al. 2007). There is also a link

between socio-economic status (Cherkas et al. 2006), obesity and related variables

(Gardner et al. 2005; Valdes et al. 2005; Cherkas et al. 2008; Nettleton et al. 2008),

124

smoking (Valdes et al. 2005), vitamin D levels (Richards et al. 2007) and age-adjusted

telomere length in very large samples of humans. There appears to be a great diversity

of influential exogenous factors, although in most cases a link is proposed between the

exogenous factor and oxidative stress [and⁄or suppression of anti-oxidative capacity;

(Serra et al. 2003)] or processes that increase cellular turnover, such as tissue

inflammation, thereby providing a mechanism for effecting telomere dynamics.

Telomere length and dynamics appear to chronicle heredity, cumulative somatic cell

growth and replicative history, exposure to oxidative stress and associated influences of

exogenous processes and agents, over and above chronological age. This has led many

authors to propose that telomeres are indicative of ‘biological age’ if not chronological

age. Here, I define biological age as variation in longevity (Nakagawa et al. 2004) and

age-relative physiological condition among individuals and species. A number of studies

on both wild animals (Haussmann et al. 2005; Pauliny et al. 2006; Bize et al. 2009;

Salomons et al. 2009) and humans (Cawthon et al. 2003) have demonstrated that

telomere length and rates of telomere loss are predictive of life expectancy (but

seeMartin-Ruiz et al. 2005; Bischoff et al. 2006 for humans). While the notion of

biological age is interesting in relation to life history strategies and trade-offs in wild

animals (Monaghan and Haussmann 2006), the complex telomere⁄genetic⁄environment

associations can rarely differentiate between cause and covariation or consequence

(Aviv 2004), particularly in an ecological setting. Additionally, reconciling the biological

age concept with chorological age and its practical use in ecological investigations of

wild animals may be difficult. Many uses of age estimates are for age structure

information, terms of elapsed time, indicators of sexual maturity status or ‘life

experience’. All of which, by definition, cannot be accurately indicated by measures of

biological age.

Determining telomere–age relationships has predominately involved evaluating cross-

sectional relationships in populations. These studies are population specific and

generally constrained to a single time point. The relationships derived are inevitably

dependent on the coverage and characteristics of the starting age distribution and so

are influenced by sampling error and⁄or bias. The method advocated for age estimation

via telomere analysis was a development of a telomere length on age ‘calibration’ for a

particular species or more recently, population (Vleck et al. 2003; Haussmann and

Mauck 2008a). Unfortunately, there have been few statistics offered in many of these

studies that allow appraisal of the error of the technique for individual age estimates -

the coefficient of determination (r2 value), while being related to predictive confidence,

in itself does not allow this (Dapson 1980). Assuming that a representative sample is

initially collected, this rationale also assumes that the mechanisms acting to produce the

telomere length relationships are homogeneous among individuals in the species or

population and temporally. This is often not the case, particularly if applying a

‘calibration’ from captive known history populations to a wild population (e.g. Bjorndal

and Zug 1995). In a wild population setting, homogeneity in the host of variables

affecting telomere dynamics, other than age, is implausible. Deriving a significant cross-

sectional relationship between telomere length and age should therefore be regarded as

a starting point, not the end point, of investigating this age estimation technique. A

125

more appropriate end point would be an understanding of what influences this

relationship spatially and temporally, providing an understanding of the error of the

technique when applying it to unknown samples. These are intriguing ecological

questions in themselves, given the links between telomere dynamics and individual

condition, survival and fitness that have been identified.

Investigations into Leach’s storm petrel (Oceanodroma leucorhoa) telomere dynamics

provide a good illustration of the dangers of interpreting cross-sectional data in an

individual framework. The telomere rate of change in this relatively long-lived species

showed a positive relationship between telomere length and age, adding 75 ± 10 (SE)

base pairs per year (Haussmann et al. 2003b). This suggested that telomeres

‘lengthened’ considerably over time in this species (Vleck et al. 2003; Nakagawa et al.

2004; Juola et al. 2006 and others). However, Haussmann and Mauck (2008b) examined

variation in initial telomere length in a very large sample of newly hatched chicks and

compared that to old adults. They compared mutually exclusive mechanistic hypotheses

of the observed telomere ‘lengthening’ and concluded that long-lived individuals start

with long telomeres and variation in telomere length decreases with age as those with

short telomeres are selected against for survival to old age (see Fig. 2). Similarly,

Halaschek-Wiener et al. (2008) have demonstrated reduced variation in telomere length

in very old humans, and Salomons et al. (2009) have demonstrated selective

disappearance of individuals with short telomeres. These intriguing observations are

nonetheless not immediately reconcilable with other observations at the individual level.

For example, individuals can actually increase telomere length over time (Aviv et al.

2008; Bize et al. 2009; Hall et al. 2004; Nordfjall et al. 2009; Ujvari & Madsen 2009;

Chapter 5) and those with the longest telomeres also appear to show the highest rates

of telomere attrition (Hall et al. 2004; Aviv et al. 2008; Bize et al. 2009; Nordfjall et al.

2009). There are two points to take away from these studies: (i) it is impossible to imply

individual level processes from small cross-sectional samples, and (ii) using the original

cross-sectional relationship as the ‘calibration’ to use for age estimates of unknown

individuals would produce erroneous individual age estimates given the proposed

processes at work (Fig. 2).

Species suitable for telomere-based chronological age estimation are those that meet

two criteria; a large change in telomere length over their lifetime and low variation in

initial telomere length (Nakagawa et al. 2004). Given that the rate of telomere change

decreases with increasing lifespan (Haussmann et al. 2003b), it is probable that species

that meet the first criterion will only be those with relatively short lifespans. However,

the second criterion may prove to be more problematic. There appears to be

considerable variation in telomere length for any one age. For example, telomere

metrics from hatchling Leach’s storm petrels vary between 3 and 4 kb (dependent of

telomere size classes measured): the same order of variation shown across the entire

age distribution in earlier cross-sectional studies in this species (Haussmann et al. 2003b).

The level of variation in initial telomere length and dynamics is likely to be species, and

given parental processes, cohort specific. Indications from some short-lived mammals

(Coviello-McLaughlin and Prowse 1997), teleosts (Ying 2005; Hatakeyama et al. 2008)

and bird species (Haussmann et al. 2003b) indicate generally high intra-age class

126

telomere variation, which is to be expected given the host of factors influencing initial

telomere length and subsequent dynamics.

In this way, the hypotheses that telomere length is a marker of chronological age and of

‘biological age’ are conflicting. If telomere length reflects ‘biological age’, the well-

documented effects of early life processes (e.g. parental condition, provisioning and

resource acquisition) on growth, condition and survival would be reflected in telomere

dynamics. Evidence is mounting that this is the case (see above) and if so chronological

age estimates from telomere length alone are, by definition of biological age,

confounded by condition.

There appear to be a few successful applications of using telomere length to estimate

individual age. However, a taxon may display correlations between telomere length and

age, particularly when telomere metrics are obtained from TRF style assays. It is clear

that, given the host of factors affecting telomere dynamics, telomeres cannot simply be

viewed as a ‘molecular clock’. There is large intra-individual variation in telomere length

at any given age and therefore, in taxa where telomere length declines with age, it is not

apparent whether an individual with comparatively short telomeres is biologically, or

chronologically, old.

Telomere length may be useful at a population scale for comparison of relative age

structures. Many endogenous and exogenous variables can affect telomere length, but

these may average out at larger scales. However, for this approach to be valid it must be

assumed that the sum of the effects of endogenous and exogenous influences is not

significantly different between the units of comparison and over time. There does not

appear to be any reason to assume a priori that this will be the case without further

evidence, given the many population-based genetic and ecological differences in species.

It would also need to be assumed that the data used to parameterise models is

representative of the spectrum of variation of influential endogenous and exogenous

factors over time. If telomere length distributions between populations are used as a

supporting basis for such studies, the influences of ecological⁄genetic processes and⁄or

differences must be accounted for. Despite recognition of the large variation in telomere

length and dynamics in individuals, the most influential factors on this variation are still

barely understood for the most intensively studied group of animals: humans. It

therefore seems unlikely that these influential factors can be accounted for robustly in

wild animal populations at present.

Current evidence suggests that telomere length alone should not be used to predict

chronological age. More promising approaches are likely to involve identification and

inclusion of influential covariates to telomere dynamics in predictive models. However,

more basic information is needed to expand and properly evaluate this. I recommend

three broad areas for future work. The first is the nature of variability in telomere length

at any given age and the processes affecting this. In the context of age estimation, this

will require assessing variation in telomere length within age classes from large samples

and across cohorts. It will also require a more sound understanding of both genetic and

environmental factors which contribute to telomere length and dynamics, best informed

127

by experimental manipulations in addition to following individuals over time exposed to

different resource acquisition regimes and other relevant environmental conditions.

The second is whether the age resolution obtained is sufficient to be used for

downstream applications. This will depend on the relationships derived, their associated

errors, an understanding of what affects them and the scale of application of the data.

However, even if a resolution of only approximately one-third of a lifespan can be

obtained (and even this with individual outliers; see above) (Takasaki et al. 2003; Vleck

et al. 2003), this is clearly insufficient for calculating terms of elapsed time, judging

sexual maturity status, individual experience and quantitative applications such as age

structure information for population dynamics assessment. If, however, one wishes to

differentiate between ‘mature’ and ‘very old’, it may be that telomere length could

provide this information if the nature of variability and processes affecting it are

understood.

Of course, if telomere length yields a sufficiently strong robust relationship for age and

its variability is well established and understood, other applications are feasible.

Currently, studies have not provided enough information to evaluate this, most that

have concluded telomeres are unfavourable for most age data applications, and there is

mounting evidence to suggest that a single telomere length age relationship should not

be applied to unknown individuals. Finally, regardless of the complexity of the models

fitted, they must be validated using known age individuals not used to parameterise

them, like other age estimation methods. The one telomere-based age estimation study

that has performed this analysis [for humans; (Takasaki et al. 2003)] had a standard

error for age estimates of 7.5 years (and therefore a 95% C. I. = 14.7 years). This study

found that 20% of estimates from a small blind sample (3⁄15) were outside of the

standard error of the technique and 13% (2⁄15) were outside of the 95% confidence

limits.

Correlations exist between telomere length and age in many taxa. Examining telomere

dynamics may lead to profound ecological and evolutionary insights, but these must be

examined and interpreted with rigour and in the right ecological context. Some basic

mechanisms of telomere dynamics are understood; however, the influences,

interactions and variation in influential endogenous and exogenous factors on these

mechanisms are currently less clear, particularly in wild animal populations. There is a

limitation on using cross-sectional sampling methods alone to shed light on these

relationships. For age estimation techniques to be successful, there is a need to have an

understanding of the error of age estimates derived from the marker under study and

how this may vary when applied to unknown individuals. This has not been made

apparent through studies undertaken thus far that promote the use of telomeres for age

estimation. The appropriateness of methods also needs to be assessed and standardized

given the nature of the samples collected and their potential genomic characteristics.

Given the conflicting nature of the biological and chronological age concepts and the

fact that these remain disparate for most practical uses of age estimates in population

ecology, telomeres are currently of limited use for determining accurate and precise

chronological age in most situations. However, the application of complex statistical

128

modelling frameworks that can incorporate influential covariates and prior information,

such as Bayesian approaches, will likely be more fruitful for future studies than the

regression based techniques used previously. Such models will require validation and

assessment as to the scale (i.e. population or individual) of their applicability and should

not be assumed to be valid for prediction because they have been made. Telomeres may

yet prove useful as age markers in some short-lived species; even so, more work needs

to be performed to understand the nature of variability in initial telomere length and

telomere dynamics. The best information about this will probably come from integrated

cross-sectional and longitudinal studies from populations under different environmental

conditions.

6.3 Conclusions

Molecular genetic techniques have revolutionized ecological and evolutionary studies

over the past three decades. Early studies predominately focused on genetic diversity

and inference of historical processes that influenced contemporary diversity within and

among species (e.g. phylogenetic and phylogeography studies). It was soon realised that

these same molecular tools could be applied over shorter timescales to illuminate

previously intractable aspects of a species’ ecology. For example, the genetic

identification of individuals facilitated unprecedented insight into parentage, kin groups,

breeding success, movement patterns and population size estimation of wild animals, to

name a few applications. This study has developed and appraised molecular genetic

tools for investigating diet and age in live marine mammals: some of the most basic of

aspects of the ecology of any species. Although results within this study were mixed

regarding specific technique efficacy, these results provide a basis for further work. Just

as this study relied upon previous conceptual advances in the understanding the genetic

patterns in species boundaries and the genetic mechanisms underlying development, it

was facilitated by technological advances in the manipulation and analysis of DNA. With

the rapid growth of molecular genetic technologies and accompanying evolving genetic

techniques, promising existing ecological genetics approaches, such as some of those in

this thesis, will be strengthened and developed further. The application of next

generation sequencing to DNA-based diet studies is already allowing larger scale studies

to be undertaken than was previously possible (e.g. Deagle et al. 2009). In addition to

strengthening previous approaches, they will also undoubtedly deliver new

opportunities. The number of genome enabled species has risen dramatically since the

human genome was published, which could eventually facilitate development of

chromosome specific telomere assays in non-model species. This would allow

unparalleled insight into telomere biology and ecology potentially aiding applications of

telomeres for age estimation, among other things. It certainly is an exciting time to be

involved in molecular ecology research.

129

References Allen, S. J., L. Bejder and M. Krutzen (2011). "Why do indo-pacific bottlenose dolphins (Tursiops sp.) carry conch shells (Tubinella sp.) in Shark Bay, Western Australia." Marine Mammal Science 27(2): 449-454. Altschul, S. F., W. Gish, W. Miller, E. W. Myers and D. J. Lipman (1997). "Basic local alignment search tool." Journal of Molecular Biology 215: 403-410. Andrew, T., A. Aviv, M. Falchi, G. L. Surdulescu, J. P. Gardner, X. Lu, M. Kimura, B. S. Kato, A. M. Valdes and T. D. Spector (2006). "Mapping genetic loci that determine leukocyte telomere length in a large sample of unselected female sibling pairs." The American Journal of Human Genetics 78: 480-486. Arisotle (≈350BC). History of Animals (Books I & II).Translated by D'arcy Wentworth Thompson.Classics.mit.edu/Aristole/history_anim.html. Arnason, U., P. W. Allderdice, J. Lien and B. Widegren (1988). "Highly Repetitive DNA in the Baleen Whale Genera Balaenoptera and Megaptera." Journal of Molecular Evolution 27: 217-221. Aviv, A. (2004). "Telomeres and Human Aging: Facts and Fibs." Science of Aging Knowledge Environment 51: pe43. Aviv, A. (2008). "The epidemiology of human telomeres: faults and promises." The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences 63: 979-983. Aviv, A., W. Chen, J. P. Gardner, M. Kimura, M. Brimacombe, X. Cao, S. R. Srinivasan and G. S. Berenson (2008). "Leukocyte Telomere Dynamics: Longitudinal Findings Among Young Adults in the Bogalusa Heart Study." Amercian Journal of Epidemiology 169(3): 323-329. Aznar, F. J., J. Hernández-Orts, A. A. Suárez, G. García-Varela, J. A. Raga and H. I. Cappozzo (2011). "Assessing host–parasite specificity through coprological analysis: a case study with species of Corynosoma (Acanthocephala: Polymorphidae) from marine mammals." Journal of Helminthology: DOI: 10.1017/S0022149X11000149. Baker, C. S., A. Perry, J. L. Bannister, M. T. Weinrich, R. B. Abernethy, J. Calambokidis, J. Lien, R. H. Lambertsen, J. U. Ramirez and O. Vasquez (1993). "Abundant mitochondrial DNA variation and world-wide population structure in humpback whales." Proceedings of the National Academy of Sciences of the United States Of America 90(17): 8239-8243. Barros, N. B. (1993) Feeding ecology and foraging strategies of bottlenose dolphins on the central east coast of Florida. PhD Dissertation. University of Miami. Coral Gables, Florida. 328 pp Barros, N. B. and M. R. Clarke (2002). Diet. Encyclopedia of Marine Mammals. W. F. Perrin, Wursig, B., Thewissen, J. G. M. San Diego, Academic Press: 323-327.

130

Barros, N. B. and D. K. Odell (1990). Food habits of bottlenose dolphins in the southeastern United States.In: The Bottlenose Dolphin.S. Leatherwood and R. R. Reeves Academic Press.San Diego. Barros, N. B., P. H. Ostrom, C. A. Stricker and R. S. Wells (2010). "Stable isotopes differentiate bottlenose dolphins off west-central Florida." Marine Mammal Science 26(2): 324-336. Barros, N. B. and R. S. Wells (1998). "Prey and feeding patterns of resident bottlenose dolphins (Tursiops truncatus) in Sarasota Bay, Florida." Journal of Mammalogy 79(3): 1045-1059. Bensasson, D., D. Zhang, D. L. Hartl and G. M. Hewitt (2001a). "Mitochondrial pseudogenes: evolutions's misplaced witnesses." Trends in ecology and evolution 16(6): 314-321. Bensasson, D., D. X. Zhang, D. L. Hartl and G. M. Hewitt (2001b). "Mitochondrial pseudogenes: evolution's misplaced witnesses." Trends In Ecology & Evolution 16(6): 314-321. Berens McCabe, E. B., D. P. Gannon, N. B. Barros and R. S. Wells (2010). "Prey selection by resident common bottlenose dolphins (Tursiops truncatus) in Sarasota Bay, Florida." Marine Biology 157: 931-942. Best, P. B. and D. M. Schell (1996). "Stable isotopes in southern right whale (Eubalaena australis) baleen as indicators of seasonal movements, feeding and growth." Marine Biology 124: 483-494. Bischoff, C., J. Graakjaer, H. C. Peterson, J. V. B. Hjelmborg, J. W. Vaupel, V. A. Bohr, S. Kølvraa and K. Christensen (2005). "The heritability of Telomere length among the elderly and oldest-old." Twin Research and Human Genetics 8(5): 433-439. Bischoff, C., H. C. Peterson, J. Graakjaer, K. Andersen-Ranberg, J. W. Vaupel, V. A. Bohr, S. Kolvraa and K. Christensen (2006). "No association between telomere length and survival among the elderly and oldest old." Epidemiology 17(2): 190-194. Bize, P., F. Criscuolo, N. B. Metcalfe, L. Nasir and P. Monaghan (2009). "Telomere dynamics rather than age predict life expectancy in the wild." Proceedings of the Royal Society B-Biological Sciences: doi: 10.1098/rspb.2008.1817. Bjorndal, K. A. and G. R. Zug (1995). Growth and age of sea turtles. Biology and conservation of sea turtles, revised edn. K. A. Bjorndal. Washington, Smithsonian Institution Press. Blackburn, E. H. (1991). "Structure and function of telomeres." Nature 350: 569-573. Blackburn, E. H. (2000). "Telomere states and cell fates." Nature 408: 53-56. Blackburn, E. H. and P. B. Challoner (1984). "Identification of a telomeric DNA sequence in Trypanosoma brucei." Cell 36(2): 447-457.

131

Blankenship, L. E. and A. A. Yayanos (2005). "Universal primers and PCR of gut contents to study marine invertebrate diets." Molecular Ecology 14(3): 891-899. Bohmann, K., A. Monadjem, C. L. Noer, M. Rasmussen, M. R. K. Zeale, E. Clare, G. Jones, E. Willerslev and M. T. P. Gilbert (2010). "Molecular Diet Analysis of Two African Free-Tailed Bats (Molossidae) Using High Throughput Sequencing." PLoS One 6(6): e21441. doi:10.1371/journal.pone.0021441. Bowles, E., P. M. Schulte, D. J. Tollit, B. E. Deagle and A. W. Trites (2011). "Proportion of prey consumed can be determined from faecal DNA using real-time PCR." Molecular Ecology Resources 11: 530-540. Boyd, I. L. (2002). "Estimating food consumption of marine predators: Antarctic fur seals and macaroni penguins." Journal of Applied Ecology 39: 103-119. Bradshaw, C. J. A., M. Hindell, A., N. J. Best, K. L. Phillips, G. Wilson and P. D. Nichols (2003). "You are what you eat: describing the foraging ecology of southern elephant seals (Mirounga leonina) using blubber fatty acids." Proceedings of the Royal Society of London. B Series 270: 1283-1292. Broccoli, D. (2004). "Function, replication and structure if the mammalian telomere." Cytotechnology 45: 3-12. Budge, S. M., S. J. Iverson and H. N. Koopman (2006). "Studying trophic ecology in marine ecosystems using fatty acids: A primer on analysis and interpretation." Marine Mammal Science 22(4): 759-801. Burk, A., J. P. Douzery and M. S. Springer (2002). "The Secondary Structure of Mammalian Mitochondrial16S rRNA Molecules: Refinements Based on a Comparative Phylogenetic Approach." Journal of Mammalian Evolution 9(3): 225-252. Campana, S. E. (2001). "Accuracy, precision and quality control in age determination, including a review of the use and abuse of age validation methods." Journal of Fish Biology 59: 197-242. Campbell, R. A., N. J. Gales, G. M. Lento and C. S. Baker (2008). "Islands in the sea: extreme female natal site fidelity in the Australian sea lion, Neophoca cinerea." Biology Letters 4(1): 139-142. Casper, R. M., S. N. Jarman, B. E. Deagle, N. J. Gales and M. Hindell, A. (2007a). "Detecting prey from DNA in predator scats: A comparison with morphological analysis, using Arctocephalus seals fed a known diet." Journal Of Experimental Marine Biology And Ecology: doi:10.1016/j.jembe.2007.04.002. Casper, R. M., S. N. Jarman, N. J. Gales and M. A. Hindell (2007b). "Combining DNA and morphological analyses of faecal samples improves insight into trophic interactions: a case study using a generalist predator." Marine Biology 152(4): 815-825. Caughley, G. (1977). Analysis of Vertebrate Populations John Wiley & Sons.London. Cawthon, R. M. (2002). "Telomere measurement by quantitative PCR." Nucleic Acids Research 30(10): e47.

132

Cawthon, R. M., K. R. Smith, E. O'Brien, A. Sivatchenko and R. A. Kerber (2003). "Association between telomere length in blood and mortality in people aged 60 years and older." Lancet 361: 393-395. Chao, A. and S. M. Lee (1992). "Estimating the number of classes via sampling coverage." Journal of the American Statistical Association 87: 210-217. Charlesworth, B., P. Sniegowski and W. Stephan (1994). "The evolutionary dynamics of repetitive DNA in eukaryotes." Nature 317(6494): 215-220. Cherkas, L. F., A. Aviv, A. M. Valdes, J. L. Hunkin, J. P. Gardner, G. L. Surdulescu, M. Kimura and T. D. Spector (2006). "The effects of social status on biological aging as measured by white-blood-cell telomere length." Aging Cell 5: 361-365. Cherkas, L. F., J. L. Hunkin, B. S. Kato, J. B. Richards, J. P. Gardner, G. L. Surdulescu, M. Kimura, X. Lu, T. D. Spector and A. Aviv (2008). "The Association Between Physical Activity in Leisure Time and Leukocyte Telomere Length." Archives of Internal Medicine 168(2): 154-158. Childerhouse, S., G. Dickie and H. G (2004). "Ageing live New Zealand sea lions (Phocarctos hookeri) using the first post-canine tooth." Wildlife Research 31(2): 177-181. Clapham, P. J., S. B. Young and R. L. Brownell Jr (1999). "Baleen whales: conservation issues and the status of the most endangered populations." Mammal Review 29(1): 35-60. Clarke, M. R. and N. Macleod (1976). "Cephalopod remains fomr sperm whales caught off Iceland." Journal of the marine biology association of the United Kingdom 56: 733-749. Cobb, B. D. and J. M. Clarkson (1994). "A simple procedure for optimising the polymerase chain reaction (PCR) using modified Taguchi methods." Nucleic Acids Research 22(18): 3801-3805. Collura, R. V. and C. Stewart (1995). "Insertions and duplications of mtDNA in the nuclear genomes of Old World monkeys and hominoids." Nature 378: 485-489. Connor, R. C., M. R. Heithaus, P. Berggren and J. L. Miksis (2000). "'Kerplunking': Surface fluke-splashes during shallow water bottom foraging by bottlenose dolphins." Marine Mammal Science 16: 646-653. Cook, P. E., L. E. Hugo, I. Iturbe-Omaetxe, C. R. Williams, S. F. Chenoweth, S. A. Ritchie, P. A. Ryan, B. H. Kay, M. W. Blows and J. P. O'Neill (2006). "The use of transcriptional profiles to predict adult mosquito age under field conditions." Proceedings of the National Academy of Sciences of the United States of America 103(48): 18060 - 18065. Cook, P. E., L. E. Hugo, I. Iturbe-Omaetxe, C. R. Williams, S. F. Chenoweth, S. A. Ritchie, P. A. Ryan, B. H. Kay, M. W. Blows and S. L. O'Neill (2007). "Predicting the age of mosquitoes using transcriptional profiles." Natuer Protocols 2(11): 2796 - 2806.

133

Coviello-McLaughlin, G. M. and K. R. Prowse (1997). "Telomere length regulation during postnatal development and ageing in Mus spretus." Nucleic Acids Research 25: 3051-3058. Criscuolo, F., P. Bize, L. Nasir, N. B. Metcalfe, C. G. Foote, K. Griffiths, E. A. Gault and P. Monaghan (2009). "Real-time quantitative PCR assay for measurement of avian telomeres." Journal of Avian Biology 40: 342-347. Dapson, R. W. (1980). "Guidelines for statistical usage in age-estimation technics." Journal of Wildlife Managment 44(3): 541-548. Darwin, C. (2003). The Origin of Species.Signet Classic.New York. de la Mare, W. K. (1987). "Fitting population models to combinations of CPUE series and absolute abundance estimates." Report to the International Whaling Commission 37: 379-381. Deagle, B. E., A. Chiaradia, J. McInnes and S. N. Jarman (2010). "Pyrosequencing faecal DNA to determine diet of little penguins: is what goes in what comes out?" Conservation Genetics 11: 2039-2048. Deagle, B. E., J. P. Eveson and S. N. Jarman (2006). "Quantification of damage in DNA recovered from highly degraded samples – a case study on DNA in faeces." Frontiers in Zoology 3(11). Deagle, B. E., N. J. Gales, K. Evans, S. N. Jarman, S. M. Robinson, R. Trebilco and M. Hindell, A. (2007). "Studying seabird diet through genetic analysis of faeces: A case study on Macaroni Penguins (Eudyptes chrysolophus)." PLoS One 2(9): e831. doi:10.1371/journal.pone.0000831. Deagle, B. E., S. N. Jarman, D. Pemberton and N. J. Gales (2005a). "Genetic screening for prey in the gut contents from a giant squid (Architeuthis sp.)." Journal Of Heredity 96(4): 417-423. Deagle, B. E., R. Kirkwood and S. N. Jarman (2009). "Analysis of Australian fur seal diet by pyrosequencing prey DNA in faeces." Molecular Ecology. Deagle, B. E., R. Kirkwood and S. N. Jarman (In Press). "Analysis of Australian fur seal diet by pyrosequencing prey DNA in faeces." Molecular Ecology. Deagle, B. E. and D. J. Tollit (2006). "Quantitative analysis of prey DNA in pinniped faeces: potential to estimate diet composition?" Conservation Genetics: DOI 10.1007/s10592-00609197-7. Deagle, B. E., D. J. Tollit, S. N. Jarman, M. A. Hindell, A. W. Trites and N. J. Gales (2005b). "Molecular scatology as a tool to study diet: analysis of prey DNA in scats from captive Steller sea lions." Molecular Ecology 14(6): 1831-1842. Delany, M. E., A. B. Krupkin and M. M. Miller (2000). "Organization of telomere sequences in birds: Evidence for arrays of extreme length and for in vivo shortening." Cytogenetics and Cell Genetics 90: 139-145.

134

Dmitriev, P. V. and Y. S. Vassetzky (2009). "Analysis of telomeric DNA: Current approaches and methods." Russian Journal of Developmental Biology 40: 125-144. Dunbar, J., S. M. Barns, L. M. Ticknor and C. R. Kuske (2002). "Empirical and Theoretical Bacterial Diversity in Four Arizona Soils." Applied And Environmental Microbiology 68(6): 3035-3045. Edgar, R. C. (2004). "MUSCLE: multiple sequence alignment with high accuracy and high throughput." Nucleic Acids Research 32(5): 1792-1797. Efron, B. and R. J. Tibshirani (1993). An introduction to the Bootstrap CRC Press LLC.Boca Raton, Florida, USA. Epel, E. S., E. H. Blackburn, J. Lin, F. S. Dhabhar, N. E. Adler, J. D. Morrow and R. M. Cawthon (2004). "Accelerated telomere shortening in response to life stress." Proceedings of the National Academy of Sciences of the United States of America 101(49): 17312-17315. Evans, K. and M. Hindell, A. (2004). "The diet of sperm whales (Physeter macrocephalus) in southern Australian waters." ICES Journal of Marine Science 61(8): 1313-1329. Fiesta-Bianchet, M., P. Blanchard, J. M. Gaillard and J. M. Hewison (2002). "Tooth extraction is not an acceptable technique to age live ungulates." Wildlife Society Bulletin 30(1): 282-283. Flinn, R., A. W. Trites, E. J. Gregr and R. I. Perry (2002). "Diets of fin, sei, and sperm whales in British Columbia: an analysis of commercial whaling records, 1963 - 1967." Marine Mammal Science 18: 663-679. Frie, A. K., V. A. Potelov, M. C. S. Kingsley and T. Haug (2003). "Trends in age-at-maturity and growth parameters of female Northeast Atlantic harp seals, Pagophilus groenlandicus (Erxleben, 1777)." Ices Journal Of Marine Science 60(5): 1018-1032. Friedrich, U., E. Griese, M. Schwab, P. Fritz, K. Thon and U. Klotz (2000). "Telomere length in different tissues of elderly patients." Mechanisms of Ageing and Development 119: 89-99. Gales, N. J., T. Kasuya, P. J. Clapham and R. L. Brownell Jr (2005). "Japan's whaling plan under scrutiny." Nature 435: 883-884. Gannon, D. P., N. B. Barros, D. P. Nowacek, A. J. Read, D. M. Waples and R. S. Wells (2005). "Prey detection by bottlenose dolphins, Tusiops truncatus: an experimental test of the passive listening hypothesis." Animal Behaviour 69: 709-720. Garde, E., A. K. Frie, G. J. Dunshea, S. H. Hansen, K. M. Kovacs and C. Lydersen (2010). "Harp seal ageing techniques-teeth, aspartic acid racemization, and telomere sequence analysis." Journal of Mammalogy 91: 1365-1374. Garde, E., M. P. Heide-Jørgensen, S. H. Hansen, G. Nachman and M. C. Forchhammer (2007). "Age-specific growth and remarkable longevity in narwhals (Monodon monoceros) from West Greenland as estimated by aspartic acids racemization." Journal of Mammology 88(1): 49-58.

135

Gardner, J. P., S. Li, S. R. Srinivasan, W. Chen, M. Kimura, X. Lu, G. S. Berenson and A. Aviv (2005). "Rise in Insulin Resistance Is Associated With Escalated Telomere Attrition." Circulation 111: 2171-2177. George, J. C., J. Bada, J. Zeh, L. Scott, S. E. Brown, T. O'Hara and R. Suydam (1999). "Age and growth estimates of bowhead whales (Balaena mysticetus) via aspartic acid racemization." Canadian Journal of Zoology 77: 571-580. Gonzalez, J. M., J. Zimmermann and C. Saiz-Jimenez (2004). "Evaluating putative chimeric sequences from PCR amplified products." Bioinformatics 21(3): 333-337. Graakjaer, J., L. Pascoe, H. Der-Sarkissian, G. Thomas, S. Kolvraa, K. Christensen and J. Londoño-Vallejo (2004). "The relative lengths of individual telomeres are defined in the zygote and strictly maintained during life." Aging Cell 3: 97-102. Grahl-Nielsen, O., M. Andersen, A. E. Derocher, C. Lydersen, O. Wiig and K. M. Kovacs (2004). "Reply to Comment on Grahl-Nielsen et al. (2003): sampling, data treatment and predictions in investigations on fatty acids in marine mammals." Marine Ecology Progress Series 281: 303-306. Green, M. L., D. L. Herzing and J. D. Baldwin (2007). "Noninvasive methodology for the sampling and extraction of DNA from free-ranging Atlantic spotted dolphins (Stenella frontalis)." Molecular Ecology Notes 7(6): 1287-1292. Green, S. J. and D. Minz (2005). "Suicide polymerase endonuclease restriction, a novel technique for enhancing PCR amplification of minor DNA templates." Applied And Environmental Microbiology 71(8): 4721-4727. Greider, C. W. (1999). "Telomeres Do D-Loop-T-Loop." Cell 97: 419-422. Gygax, L. (2002). "Evolution of group size in the superfamily Delphinoidea (Delphinidae, Phocoenidae and Monodontidae): a quantitative comparative analysis." Mammal Review 32(4): 295-314. Haas, B. J., D. Gevers, A. M. Earl, M. Feldgarden, D. V. Ward, G. Giannoukos, D. Ciulla, D. Tabbaa, S. K. Highlander, E. Sodergren, B. Methe, T. Z. DeSantis, J. F. Petrosino, R. Knight and B. W. Birren (2011). "Chimeric 16S rRNA sequence formation and detection in Sanger and 454-pyrosequenced PCR amplicons." Genome Research 21(3): 494-504. Hajibabaei, M., D. H. Janzen, J. M. Burns, W. Hallwachs and P. D. N. Hebert (2006). "DNA barcodes distinguish species of tropical Lepidoptera." Proceedings Of The National Academy Of Sciences Of The United States Of America 103(4): 968-971. Hall, M. E., L. Nasir, F. Daunt, E. A. Gault, J. P. Croxall, S. Wanless and P. Monaghan (2004). "Telomere loss in relation to age and early environment in long-lived birds." Proceedings of the Royal Society of London. B Series 271: 1571-1576. Hall, T. A. (1999). "BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT." Nucleic Acids Symposium Series 41: 95-98.

136

Harcourt, R. G., J. J. Kingston, M. Cameron, J. R. Waas and M. Hindell, A. (2007). "Paternity analysis shows experience, not age, enhances mating success in an aquatically mating pinniped, the Weddell seal (Leptonychotes weddellii)." Behavioral Ecology and Sociobiology 61: 643-652. Harley, C. B., A. B. Futcher and C. W. Greider (1990). "Telomeres shorten during ageing of human fibroblasts." Nature 345: 458-460. Harper, G. L., R. A. King, C. S. Dodd, J. D. Harwood, D. M. Glen, M. W. Bruford and W. O. C. Symondson (2005). "Rapid screening of invertebrate predators for multiple prey DNA targets." Molecular Ecology 14(3): 819-827. Hastie, N. D., M. Dempster, M. G. Dunlop, A. M. Thompson, D. K. Green and R. C. Allshire (1990). "Telomere reduction in human colorectal carcinoma and with ageing." Nature 346: 866-868. Hatakeyama, H., K. I. Nakamura, N. Izumiyama-Shimorriura, A. Ishii, S. Tsuchida, K. Takubo and N. Ishikawa (2008). "The teleost Oryzias latipes shows telomere shortening with age despite considerable telomerase activity throughout life (vol 129, pg 550, 2008)." Mechanisms Of Ageing And Development 129(11): 692-692. Hatase, H., R. Sudo, K. K. Watanabe, T. Kasugai, T. Saito, H. Okamoto, I. Uchida and K. Tsukamoto (2008). "Shorter telomere length with age in the loggerhead turtle: a new hope for live sea turtle age estimation." Genes and Genetic Systems 83: 423-426. Haussmann, M. F. and R. A. Mauck (2008a). "New strategies for telomere-based age estimation." Molecular Ecology Notes 8(2): 264-274. Haussmann, M. F. and R. A. Mauck (2008b). "Telomeres and longevity: Testing an evolutionary hypothesis." Molecular Biology and Evolution 25: 220-228. Haussmann, M. F. and C. M. Vleck (2002). "Telomere length provides a new technique for aging animals." Oecologia 130: 325-328. Haussmann, M. F., C. M. Vleck and I. C. T. Nisbet (2003a). "Calibrating the telomere clock in common terns, Sterna hirundo." Experimental Gerontology 38: 787-789. Haussmann, M. F., D. W. Winkler, K. M. O'Reilly, C. E. Huntington, I. C. T. Nisbet and C. M. Vleck (2003b). "Telomeres shorten more slowly in long-lived birds and mammals than in short-lived ones." Proceedings of the Royal Society of London. B Series 270(1522): 1387-1392. Haussmann, M. F., D. W. Winkler and C. M. Vleck (2005). "Longer telomeres associated with higher survival in birds." Biology Letters 1: 212-214. Heithaus, M. R. (2001). "Predator-prey and competitive interactions between sharks (order Selachii) and dolphins (suborder Odontoceti): a review." Journal of Zoology 253: 53-68. Henmann, M. T., M. A. Strong, L. Y. Hao and C. W. Grieder (2001). "The shortest telomere, not average telomere length, is critical for cell viability and chromosome stability." Cell 107: 67-77.

137

Herman, D. P., C. O. Matkin, G. M. Ylitalo, J. W. Durban, M. Bradley Hanson, M. E. Dahlheim, J. M. Straley, P. R. Wade, K. L. Tilbury, R. H. Boyer, R. W. Pearce and M. M. Krahn (2008). "Assessing age distributions of killer whale Orcinus orca populations from the composition of endogenous fatty acids in their outer blubber layers." Marine Ecology Progress Series 372(289-302). Herman, D. P., G. M. Ylitalo, R. J., J. M. Straley, C. M. Gabriele, P. J. Clapham, R. H. Boyer, K. L. Tilbury, R. W. Pearce and M. M. Krahn (2009). "Age determination of humpback whales Megaptera novaeangliae through blubber fatty acid compositions of biopsy samples." Marine Ecology Progress Series 392: 277-293. Hoelzel, A. R., J. Hey, M. E. Dahlheim, C. Nicholson, V. Burkanov and N. Black (2007). "Evolution of Population Structure in a Highly Social Top Predator, the Killer Whale." Molecular Biology and Evolution 24(6): 1407-1415. Hoffman, J. I., C. W. Matson, W. Amos, T. R. Loughlin and J. W. Bickham (2006). "Deep genetic subdivision within a continuously distributed and highly vagile marine mammal, the Steller’s sea lion (Eumetopias jubatus)." Molecular Ecology 15: 2821-2832. Hogg, C. J., T. L. Rogers, A. Shorter, K. Barton, P. J. O. Miller and D. Nowacek (2009). "Determination of steroid hormones in whale blow: It is possible." Marine Mammal Science 25(3): 605-618. Hohn, A. A., M. D. Scott, R. S. Wells, J. C. Sweeny and A. B. Irvine (1989). "Growth layers in teeth from known-age, free-ranging bottlenose dolphins." Marine Mammal Science 5(4): 315-342. Holmes, E. E. and A. E. York (2003). "Using Age Structure to Detect Impacts on Threatened Populations: a Case Study with Stellar Sea Lions." Conservation Biology 17(6): 1794-1806. Horn, T., B. C. Robertson and N. J. Gemmell (2010). "The use of telomere length in ecology and evolutionary biology." Heredity 105: 497-506. Huber, J. A., D. A. Butterfield and J. A. Baross (2002). "Temporal changes in archaeal diversity and chemistry in a mid-ocean ridge subseafloor habitat." Applied And Environmental Microbiology 68(4): 1585-1594. Hughes, J. B., J. J. Hellmann, T. H. Ricketts and B. J. M. Bohannan (2001). "Counting the Uncountable: Statisical Approaches to Estimating Microbial Diversity." Applied And Environmental Microbiology 67(10): 4399-4406. Iverson, S. J., C. Feild, W. D. Bowen and W. Blanchard (2004). "Quantitative fatty acid signature analysis: a new method of estimating predator diets." Ecological Monographs 74: 211-235. Iverson, S. J., A. M. Springer and A. S. Kitaysky (2007). "Seabirds as indicators of food web structure and ecosystem variability: qualitative and quantitative diet analyses using fatty acids." Marine Ecology Progress Series 352: 235-244.

138

Jansen, O. E., M. F. Leopold, E. H. W. G. Meesters and C. Smeenk (2010). "Are white-beaked dolphins Lagenorhynchus albirostris food specialists? Their diet in the southern North Sea." Journal of the Marine Biological Association of the United Kingdom 90(8): 1501-1508. Jarman, S. N. (2004). "Amplicon: software for designing PCR primers on aligned DNA sequences." Bioinformatics 20(10): 1644-1645. Jarman, S. N. (2006). "Cleaver: software for identifying taxon specific restriction endonuclease recognition sites." Bioinformatics 22: 2160-2161. Jarman, S. N., B. E. Deagle and N. J. Gales (2004). "Group-specific polymerase chain reaction for DNA-based analysis of species diversity and identity in dietary samples." Molecular Ecology 13: 1313-1322. Jarman, S. N., N. J. Gales, M. Tierney, P. C. Gill and N. G. Elliot (2002). "A DNA-based method for identification of krill species and its application to analysing the diet of marine vertebrate predators." Molecular Ecology 11: 2679-2690. Jarman, S. N., K. S. Redd and N. J. Gales (2006). "Group-specific primers for amplifying DNA sequences that identify Amphipoda, Cephalopoda, Echinodermata, Gastropoda, Isopoda, Ostracoda and Thoracica." Molecular Ecology Notes 6(1): 268-271. Jenkins, G. P. (1988). "Micro- and fine-scale distribution of microplankton in the feeding environment of larval flounder." Marine Ecology Progress Series 43: 233-244. Ju, S. J., D. H. Secor and H. R. Harvey (1999). "Growth rate variability and lipofuscin accumulation rates in the blue crab Callinectes sapidus." Marine Ecology Progress Series 224: 197-205. Juen, A. and M. Traugott (2005). "Detecting predation and scavenging by DNA gut-content analysis: a case study using a soil insect predator-prey system." Oecologia 142(3): 344-352. Juola, F. A., M. F. Haussmann, D. C. Dearborn and C. M. Vleck (2006). "Telomere shortening in a long-lived marine bird: cross-sectional analysis and test of an aging tool." The Auk 123(3): 775-783. Karlsson, A. O., A. Svensson, A. Marklund and G. Holmlund (2008). "Estimating human age in forensic samples by analysis of telomere repeats." Forensic Science International: Genetics Supplement Series 1: 569-571. Kelly, C. and T. D. Price (2005). "Correcting for regression to the mean in behavior and ecology." American Naturalist 166: 700-707. Kimura, M. and A. Aviv (2011). "Measurement of telomere DNA content by dot blot analysis." Nucleic Acids Research: doi:10.1093/nar/gkr235. Kimura, M., R. C. Stone, S. C. Hunt, J. Skurnick, X. Lu, X. Cao, C. B. Harley and A. Aviv (2010). "Measurement of telomere length by southern blot analysis of the terminal restriction fragment lengths." Nature Protocols 5: 1596-1607.

139

King, R. A., D. S. Read, M. Traugott and W. O. C. Symondson (2008). "Molecular analysis of predation: a review of best practice for DNA-based approaches." Molecular Ecology 17(4): 947-963. Kipling, D. and H. J. Cooke (1990). "Hypervariable ultra-long telomeres in mice." Nature 347: 400-402. Kirby, L. T. (1992). DNA Fingerprinting W. H. Freeman and Company.New York. Klevezal, G. A. and S. E. Kleinenberg (1969). Age determination of mammals from annual layers in teeth and bones.The U.S. Department of the Interior and The National Science Foundation.T. U. S. D. o. t. I. a. T. N. S. Foundation.Washington D. C. Kobayashi, N., K. Tamura and T. Aotsuka (1999). "PCR error and molecular population genetics." Biochemical Genetics 37: 317-321. Kohn, M. H. and R. K. Wayne (1997). "Facts from feces revisited." Trends in ecology and evolution 12: 223-227. Kotrschal, A., P. IImonen and D. J. Penn (2007). "Stress impacts telomere dynamics." Biology Letters 3(128-130). Krafft, B. A., K. M. Kovacs, A. K. Frie, T. Haug and C. Lydersen (2006). "Growth and population parameters of ringed seals (Pusa hispida) from Svalbard, Norway, 2002–2004." Ices Journal Of Marine Science 63(6): 1136-1144. Kumar, S., K. Tamura and M. Nei (2004). "MEGA3: Integrated software for Molecular Evolutionary Genetics Analysis and sequence alignment." Briefings in Bioinformatics 5: 150-163. Kvitrud, M. A., S. D. Riemer, R. F. Brown, M. R. Bellinger and M. A. Banks (2005). "Pacific harbor seals (Phoca vitulina) and salmon: genetics presents hard numbers for elucidating predator-prey dynamics." Marine Biology 147(6): 1459-1466. Lasley, B. L. and J. F. Kirkpatrick (1991). "Monitoring of ovarian function in captive and free-ranging wildlife by means of urinary and fecal steroids." Journal of Zoo and Wildlife Medicine 22(1): 23-31. Leaper, R., J. Cooke, P. Trethan, K. Reid, V. Rowntree and R. Payne (2006). "Global climate drives southern right whale (Eubalaena australis) population dynamics." Biology Letters 2: 289-292. Lesnoff, M. and R. Lancelot (2010). "aod: Analysis of Overdispersed Data." R Package version 1.2, URL http://cran.r-project.org/package=aod. Li, T., C. Kim, C. M. Rubin and C. W. Schmid (2000). "K562 cells implicate increased chromatin accessibility in Alu transcriptional activation." Nucleic Acids Research 28(16): 3031-3039. Lockyer, C. (1984). "Age determination by means of the ear plug in baleen whales." Report to the International Whaling Commission 34: 692-696.

140

Lopez, J. V., N. Yuhki, R. Masuda, W. Modi and S. J. O'brien (1994). "Numt, a recent transfer and tandem amplification of mitochondrial DNA to the nuclear genome of the domestic cat." Journal of Molecular Evolution 39(2): 174-190. Marmontel, M., T. J. Nhea, H. I. Kochman and S. R. Humphrey (1996). "Age determination in manatees using growth-layer-group counts in bone." Marine Mammal Science 12(1): 54-88. Martin-Ruiz, C. M., J. Gussekloo, D. van Heemst, T. von Zglinicki and R. G. J. Westendorp (2005). "Telomere length in white blood cells is not associated with morbidity or mortality in the oldest old: a population-based study." Aging Cell 4: 287-290. Martin, D. L., R. M. Ross, L. B. Quetin and A. E. Murray (2006). "Molecular approach (PCR-DGGE) to diet analysis in young Antarctic krill Euphausia superba." Marine Ecology Progress Series 319: 155-165. McLeod, C. D., M. B. Santos and G. J. Pierce (2003). "Review of data on diets of beaked whales: evidence of niche separation and geographic segregation." Journal of the Marine Biological Association of the United Kingdom 83: 651-665. Messier, T. L., J. P. O'Neill, S. Hou, J. A. Nicklas and B. A. Finette (2003). "In vivo transposition mediated by V(D)J recombinase in human T lymphocytes." The EMBO Journal 22: 1381-1388. Meyne, J., R. J. Baker, H. H. Hobart, T. C. Hsu, O. A. Ryder, O. G. Ward, J. E. Wiley, D. H. Wurster-Hill, T. L. Yates and R. K. Moyzis (1990). "Distribution of non-telomeric sites of the (TTAGGG)n telomeric sequence in vertebrate chromosomes." Chromosoma 99: 3-10. Mirol, P. M., S. Mascheretti and J. B. Searle (2000). "Multiple nuclear pseudogenes of mitochondrial cytochrome b in Ctenomys (Caviomorpha, Rodentia) with either great similarity to or high divergence from the true mitochondrial sequence." Heredity 84: 538-547. Monaghan, P. and M. F. Haussmann (2006). "Do telomere dynamics link lifestyle and lifespan?" Trends in ecology and evolution 21(1): 47-53. Morin, P. A., N. M. Hedrick, K. M. Robertson and C. A. Leduc (2006). "Comparative mitochondrial and nuclear quantitative PCR of historical marine mammal tissue, bone, baleen, and tooth samples." Molecular Ecology Notes 7(3): 404-411. Morin, P. A., K. K. Martien, F. I. Archer, F. Cipriano, D. Steel, J. Jackson and B. L. Taylor (2010). "Applied Conservation Genetics and the Need for Quality Control and Reporting of Genetic Data Used in Fisheries and Wildlife Management." Journal of Heredity 101(1): 1-10. Morrison, T. A., P. I. Chiyo, C. J. Moss and S. C. Alberts (2005). "Measures of dung bolus size for known-age African elephants (Loxodonta africana): implications for age estimation." Journal of Zoology 266: 89-94. Moyzis, R. K., J. M. Buckingham, L. S. Cram, M. Dani, L. L. Deaven, M. D. Jones, J. Meyne, R. L. Ratliff and J. Wu (1988). "A highly conserved repetitive DNA sequence, (TTAGGG)n,

141

present at the telomeres of human chromosomes." Proceedings of the National Academy of Sciences of the United States of America 85: 6622-6626. Munch, K., W. Boomsma, J. P. Huelsenbeck, E. Willerslev and N. Neilsen (2008a). "Statistical Assignment of DNA Sequences using Bayesian Phylogenetics." Systematic Biology 57(5): 750-757. Munch, K., W. Boomsma, E. Willerslev and N. Neilsen (2008b). "Fast phylogenetic DNA barcoding." Philosophical Transactions Of The Royal Society B-Biological Sciences. Munch, K., W. Boomsma, E. Willerslev and N. Neilsen (2008c). "Statistical Assignment of DNA Sequences using Bayesian Phylogenetics." Systematic Biology 57(5): 750-757. Nakagawa, S., N. J. Gemmell and T. Burke (2004). "Measuring vertebrate telomeres: applications and limitations." Molecular Ecology 13: 2523-2533. Nakamura, K., N. Izumiyama-Shimomura, M. Sawabe, T. Arai, Y. Aoyagi, M. Fujiwara, E. Tsuchiya, Y. Kobayashi, M. Kato, M. Oshimura, K. Sasajima, K. Nakachi and K. Takubo (2002). "Comparative Analysis of Telomere Lengths and Erosion with Age in Human Epidermis and Lingual Epithelium." Journal of Investigative Dermatology 119: 1014-1019. Nejstgaard, J. C., M. E. Frischer, P. Simonelli, C. Troedsson, M. Brakel, F. Adiyaman, A. F. Sazhin and L. F. Artigas (2008). "Quantitative PCR to estimate copepod feeding." Marine Biology 153: 565-577. Nelson, M. E. (2002). "The Science, Ethics and Philosophy of Tooth Extractions from Live-Captured White-Tailed Deer: A Response to Fiesta-Bianchet et al. (2002)." Wildlife Society Bulletin 30(1): 284-288. Nettleton, J. A., A. Diez-Roux, N. S. Jenny, A. L. Fitzpatrick and D. R. Jacobs Jr (2008). "Dietary patterns, food groups, and telomere length in the Multi-Ethnic Study of Atherosclerosis(MESA)." American Journal of Clinical Nutrition 88: 1405-1412. Newland, C., I. C. Feild, P. D. Nichols, C. J. A. Bradshaw and M. Hindell, A. (2009). "Blubber fatty acid profiles indicate dietary resource partitioning between adult and juvenile southern elephant seals." Marine Ecology Progress Series 384: 303-312. Newsome, S. D., M. R. Clementz and P. L. Koch (2010). "Using stable isotope biochemistry to study marine mammal ecology." Marine Mammal Science 26(3): 509-572. Nielsen, R. and M. Matz (2006). "Statistical approaches for DNA barcoding." Systematic Biology 55(1): 162-169. Nordfjall, K., A. Larefalk, P. Lindgren, D. Holmberg and G. Roos (2005). "Telomere length and heredity: Indications of paternal inheritance." Proceedings of the National Academy of Sciences of the United States of America 102(45): 16374-16378. Nordfjall, K., U. Svenson, K. Norrback, R. Adolfsson, P. Lenner and G. Roos (2009). "The Individual Blood Cell Telomere Attrition Rate is Telomere Length Dependent." PLoS Genetics 5(2): e1000375.doi:10.1371/journal.pgen.1000375.

142

O'Callaghan, N. J., V. S. Dhillon, P. Thomas and M. Fenech (2008). "A quantitative real-time PCR method for absolute telomere length." Biotechniques 44: 807-809. O'Corry-Crowe, G. M., R. S. Suydam, A. Rosenberg, K. J. Frost and A. E. Dizon (1997). "Phylogeography, population structure and dispersal patterns of the beluga whale Delphinapterus leucas in the western Nearctic revealed by mitochondrial DNA." Molecular Ecology 6(10): 955-970. OEHM, J., A. Juen, K. Nagiller, S. Neuhauser and M. Traugott (2011). "Molecular scatology: how to improve prey DNA detection success in avain faeces." Molecular Ecology Resources: doi:10.1111/j.1755-0998.2011.03001.x. Oftedal, O. T. (2000). "Use of maternal resreves as a lactation stragey in large mammals." Proceedings of the Nutrition Society 59: 99-106. OKazaki, S., K. Tsuchida, H. Maekawa, H. Ishikawa and H. Fujiwara (1993). "Identification of a Pentanucleotide Telomere Sequence, (TTAGG)n, in the Silkworm Bombyx mori and in Other Insects." Molecular and Cellular Biology 13(3): 1424-1432. Olovnikov, A. M. (1973). "A theory of marginotomy." Journal of Theoretical Biology 41: 181-190. Olsen, L. E. and A. D. Yoder (2002). "Using Secondary Structure to Identify Ribosomal Numts: Cautionary Examples from the Human Genome." Molecular Biology and Evolution 19(1): 93-100. Olsson, M., A. Pauliny, E. Wapstra, T. Uller, T. Schwartz and D. Blomqvist (2010). "Sex Differences in Sand Lizard Telomere Inheritance: Paternal Epigenetic Effects Increases Telomere Heritability and Offspring Survival." PLoS One 6(4): e17473. doi:10.1371/journal.pone.0017473. Olsson, M., A. Pauliny, E. Wapstra, T. Uller, T. Schwartz, E. Miller and D. Blomqvist (2011). "Sexual differences in telomere selection in the wild." Molecular Ecology 20(10): 2085-2099. Palomeque, T. and P. Lorite (2008). "Satellite DNA in insects: a review." Heredity 100: 564-573. Parsons, K. M., J. F. Dallas, D. E. Claridge, J. W. Durban, K. C. Balcomb, P. M. Thompson and L. R. Noble (1999). "Amplifying dolphin mitochondrial DNA from faecal plumes." Molecular Ecology 8(10): 1766-1768. Parsons, K. M., S. B. Piertney, S. J. Middlemas, P. S. Hammond and J. D. Armstrong (2005). "DNA-based identification of salmonid prey species in seal faeces." Journal Of Zoology 266: 275-281. Patenaude, N. J. and C. S. Baker (2001). "Population status and habitat use of southern right whales in the sub-Antarctic Auckland Islands of New Zealand." Journal of Cetacean Research & Management Special Issue.(Special Issue 2): 111-116.

143

Pauli, J. N., J. P. Whiteman, B. G. Marcot, T. M. McClean and M. Ben-David (2011). "DNA-based approach to aging martens (Martes americana and M. caurina)." Journal of Mammology 92(3): 500-510. Pauliny, A., R. H. Wagner, J. Augustin, T. Szep and D. Blomqvist (2006). "Age-independent telomere length predicts fitness in two bird species." Molecular Ecology 15: 1681-1687. Pereira, S. L. and A. J. Baker (2004). "Low number of mitochondrial pseudogenes in the chicken (Gallus gallus) nuclear genome: implications for molecular inference of population history and phylogenetics." BMC Evolutionary Biology 4(17): doi:10.1186/1471-2148-4-17. Pierce, G. J. and P. R. Boyle (1991). "A review of methods for diet analysis in picivorous marine mammals." Oceanography and Marine Biology: An Annual Review 29: 409-486. Polz, M. F. and C. M. Cavanaugh (1998). "Bias in template-to-product ratios in multitemplate PCR." Applied And Environmental Microbiology 64(10): 3724-3730. Poulakakis, N., P. Lymberakis, K. Paragamian and M. Mylonas (2005). "Isolation and amplification of shrew DNA from barn owl pellets." Biological Journal Of The Linnean Society 85(3): 331-340. Punt, A. E. (1999). "A full description of the standard BALEEN II model and some varients thereof." Journal of Cetacean Research & Management 1: 267-276. Purcell, M., G. Mackey, E. LaHood, H. Huber and L. Park (2004). "Molecular methods for the genetic identification of salmonid prey from Pacific harbor seal (Phoca vitulina richardsi) scat." Fishery Bulletin 102(1): 213-220. Reilly, J. (2002). "Growth in the Sumatran elephant (Elephas maximus sumatranus) and age estimation based on dung diameter." Journal of Zoology 258(2): 205-213. Riccialdelli, L., S. D. Newsome, M. L. Fogel and R. N. P. Goodall (2010). "Isotopic assessment of prey and habitat preferences of a cetacean community in the southwestern South Atlantic Ocean." Marine Ecology Progress Series 418: 235-248. Richards, J. B., A. M. Valdes, J. P. Gardner, D. Paximadas, M. Kimura, A. Nessa, X. Lu, G. L. Surdulescu, R. Swaminathan, T. D. Spector and A. Aviv (2007). "Higher serum vitamin D concentrations are associated with longer leukocyte telomere length in women." American Journal of Clinical Nutrition 86: 1420-1425. Richly, E. and D. Leister (2004). "NUMTs in Sequenced Eukaryotic Genomes." Molecular Biology and Evolution 21(6): 1081-1084. Ridgeway, S. H., R. F. Green and J. C. Sweeny (1975). "Mandibular anesthesia and tooth extraction in the bottlenosed dolphin." Journal of Wildlife Diseases 11: 415-418. Rodrigue, K. L., B. P. May, T. R. Famula and M. E. Delany (2005). "Meiotic instability of chicken ultra-long telomeres and mappings of a 2.8 megabase arrays to the W-sex chromosome." Chromosome Research 13(6): 581-591.

144

Rolland, R. M., K. E. Hunt, S. D. Kraus and S. K. Wasser (2005). "Assessing reproductive status of right whales (Eubalaena glacialis) using fecal hormone metabolites." General and Comparative Endocrinology 142(3): 308-317. Rosel, P. E., L. Hansen and A. A. Hohn (2009). "Restricted dispersal in a continuously distributed marine species: common bottlenose dolphins Tursiops truncatus in coastal waters of the western North Atlantic." Molecular Ecology 18(24): 5030-5045. Ross, R. A., S. Murugan and W. L. S. Li (2008). "Testing the reliability of genetic methods of species identification via simulation." Systematic Biology 57(2): 216-230. Rubenstein, D. R. and K. A. Hobson (2004). "From birds to butterflies: animal movement patterns and stable isotopes." Trends In Ecology & Evolution 19(5): 256-263. Rubinoff, D., S. Cameron and K. Will (2006). "A genomic perspective on the shrtcomings of mitochondrial DNA for "barcoding" identification." Journal Of Heredity 97(6): 581-594. Rufer, N., T. H. Brümmendorf, S. Kolvraa, C. Bischoff, K. Christensen, L. Wadsworth, M. Schulzer and P. M. Lansdorp (1999). "Telomere Fluorescence Measurements in Granulocytes and T Lymphocyte Subsets Point to a High Turnover of Hematopoietic Stem Cells and Memory T Cells in Early Childhood." Journal of Experimental Medicine 190(2): 157-167. Ruiz-Cooley, R. I., D. Gendron, S. Aguiniga, S. Mesnick and J. D. Carriquiry (2004). "Trophic relationships between sperm whales and jumbo squid using stable isotopes of C and N." Marine Ecology Progress Series 277: 275-283. Salomons, H. M., G. A. Mulder, L. van de Zande, M. F. Haussmann, M. H. K. Linskens and S. Verhulst (2009). "Telomere shortening and survival in free-living corvids." Proceedings of the Royal Society B-Biological Sciences 276: 3157-3165. Sambrook, J., E. F. Fritsch and T. Maniatis (1989). Molecular Cloning: A laboratory Manual Cold Spring Harbor Laboratory Press.Cold Spring Harbor, NY. Santos, M. B., M. R. Clarke and G. J. Pierce (2001). "Assessing the importance of cephalopods in the diets of marine mammals and other top predators: problems and solutions." Fisheries Research 52: 121-139. Schmeling, A., G. Geserick, W. Reisinger and A. Olze (2007). "Age estimation." Forensic Science International 165: 178-181. Scott, N. M., M. F. Haussmann, R. M. Elsey, P. L. Trosclair III and C. M. Vleck (2006). "Telomere length shortens with body length in Alligator mississippiensis." Southeastern Naturalist 5(4): 685-692. Serra, V., T. von Zglinicki, M. Lorenz and G. Saretzki (2003). "Extracellular superoxide dismutase is a major antioxidant in human fibroblasts and slows telomere shortening." Journal of Biological Chemistry 278: 6824-6830. Sheppard, S. K. and J. D. Harwood (2005). "Advances in molecular ecology: tracking trophic links through predator-prey food-webs." Functional Ecology 19(5): 751-762.

145

Simon, C., F. Frati, B. A and e. al. (1994). "Evolution, weighing, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers." Annals of the Entomological Society of America 87: 651-701. Slagboom, P. E., S. Droog and D. I. Boomsma (1994). "Genetic determination of telomere size in humans: a twin study of three age groups." American Journal of Human Genetics 55: 876-882. Smokler, R., A. Richards, R. C. Connor, J. Mann and P. Berggren (1997). "Sponge carrying by dolphins (Delphinidae, Tursiops sp.): A foraging specialization involving tool use?" Ethology 103: 454-465. Song, H., J. E. Buhay, M. F. Whiting and K. A. Crandall (2008). "Many species in one: DNA barcoding overestimates the number of species when nuclear mitochondrial pseudogenes are coamplified." Proceedings of the National Academy of Sciences of the United States of America 105(36): 13486-13491. Sorenson, M. D. and R. C. Fleischer (1996). "Multiple independent trans-positions of mitochondrial DNA control region sequences to the nucleus." Proceedings Of The National Academy Of Sciences Of The United States Of America 93: 15239-15243. Sorenson, M. D. and T. W. Quinn (1998). "Numts: A challenge for avian systematics and population biology." The Auk 115(1): 214-221. Springer, A., J. Estes, G. van Vliet, T. Williams, D. Doak, E. Danner and B. Pfister (2008). "Mammal-eating killer whales, industrial whaling, and the sequential megafaunal collapse in the North Pacific Ocean: A reply to critics of Springer et al. 2003." Marine Mammal Science 24(2): 414-442. Stearns, S. C. (1992). The Evolution of Life Histories Oxford University Press.Oxford. Steinke, D., M. Vences, W. Salzburger and A. Meyer (2005). "TaxI: a software tool for DNA barcoding using distance methods." Philosophical Transactions Of The Royal Society B-Biological Sciences 360(1462): 1975-1980. Stephens, D. W. and J. R. Krebs (1986). Foraging Theory Princeton University Press.Princeton. Storr, G. M. (1961). "Microscopic analysis of faeces, a technique for ascertaining the diet of herbivorous mammals." Australian Journal of Biological Sciences 14(1): 157-164. Sunnucks, P., A. C. C. Wilson, L. B. Beheregaray, K. Zenger, J. French and A. C. Taylor (2000). "SSCP is not so difficult: the application and utility of single-stranded conformational polymorphism in evolutionary biology and molecular ecology." Molecular Ecology 9: 1699-1710. Sutherland, R. M. (2000) Molecular analysis of avian diets. PhD Thesis. University of Oxford. Oxford. Symondson, W. O. C. (2002a). Diagnostic techniques for determining carabid diets. Agroecology Of Carabid Beetles. Andover, Intercept Ltd Scientific, Technical & Medical Publishers: 137-164.

146

Symondson, W. O. C. (2002b). "Molecular identification of prey in predator diets." Molecular Ecology 11(4): 627-641. Takasaki, T., A. Tsuji, N. Ikeda and M. Ohishi (2003). "Age estimation in dental pulp DNA based on human telomere shortening." International Journal of Legal Medicine 117: 232-234. Team, R. D. C. (2010). R: A language and environment for statistical computing R. F. f. S. Computing.Vienna, Austria. Thalmann, O., J. Hebler, H. N. Poinar, S. Paabo and L. Vigilant (2004a). "Unreliable mtDNA data due to nuclear insertions: a cautionary tale from analysis of humans and other great apes." Molecular Ecology 13: 321-335. Thalmann, O., J. Hebler, H. N. Poinar, S. Paabo and L. Vigilant (2004b). "Unreliable mtDNA data due to nuclear insertions: a cautionary tale from analysis of humans and other great apes." Molecular Ecology 13(2): 321-335. Thalmann, O., D. Serre, M. Hofreiter, D. Lukas, J. Eriksson and L. Vigilant (2005). "Nuclear insertions help and hinder inference of the evolutionary history of gorilla mtDNA." Molecular Ecology 14(1): 179-188. Tollit, D. J., A. D. Schulze, A. W. Trites, P. F. Olesiuk, S. J. Crockford, T. S. Gelatt, R. R. Ream and K. M. Miller (2009). "Development and application of DNA techniques for validating and improving pinniped diet estimates." Ecological Applications 19(4): 889-905. Torres, L. G. and A. J. Read (2009). "Where to catch a fish? The influence of foraging tactics on the ecology of bottlenose dolphins (Tursiops truncatus) in Florida Bay, Florida." Marine Mammal Science 25(4): 797-815. Triant, D. A. and J. A. Dewoody (2007). "The occurrence, detection, and avoidance of mitochondrial DNA translocations in mammalian systematics and phylogeography." Journal Of Mammalogy 88(4): 908-920. Trites, A. W. and R. Joy (2005). "Dietary analysis from fecal samples: How many scats are enough?" Journal of Mammalogy 86(4): 074-712. Tsuji, A., A. Ishiko, T. Takasaki and N. Ikeda (2002). "Estimating age of humans based on telomere shortening." Forensic Science International 126: 197-199. Ujvari, B. and T. Madsen (2009). "Short telomeres in Hatchling Snakes: Erythrocyte Telomere Dynamics and Longevity in Tropical Pythons." PLoS One 4(10): e7493. doi:10.1371/journal.pone.0007493. Unryn, B. M., L. S. Cook and K. T. Riabowol (2005). "Paternal age is positively linked to telomere length of children." Aging Cell 4: 97-101. Valdes, A. M., T. Andrew, J. P. Gardner, M. Kimura, E. Oelsner, L. F. Cherkas, A. Aviv and T. D. Spector (2005). "Obesity, cigarette smoking, and telomere length in women." Lancet 366: 662-64.

147

Valenzuela, L. O., M. Sironi, V. J. Rowntree and J. Seger (2009). "Isotopic and genetic evidence for culturally inherited site fidelity to feeding grounds in southern right whales (Eubalaena australis)." Molecular Ecology 18(5): 782-791. Vences, M., M. Thomas, A. van der Meijden, Y. Chiari and D. R. Vieites (2005). "Comparative performance of the 16S rDNA gene in DNA barcoding of amphibians." Frontiers in Zoology 2(5): doi:10.1186/1742-9994-2-5. Venkatesh, B., N. Dandona and S. Brenner (2006). "Fugu genome does not contain mitochondrial pseudogenes." Genomics 87(2): 307-310. Vestheim, H. and S. N. Jarman (2008). "Blocking primers to enhance PCR amplification of rare sequences in mixed samples - a case study on prey DNA in Antarctic krill stomachs." Frontiers in Zoology 5(12): doi:10.1186/1742-9994-5-12. Vitt, L. J. and E. R. Pianka (2005). "Deep history impacts present-day ecology and biodiversity." Proceedings of the National Academy of Sciences of the United States of America 102(22): 7877-7881. Vleck, C. M., M. F. Haussmann and D. Vleck (2003). "The natural history of telomeres: tools for aging animals and exploring the aging process." Experimental Gerontology 38: 791-795. von Wintzingerode, F., O. Landt, A. Ehrlich and U. B. Gobel (2000). "Peptide nucleic acid-mediated PCR clamping as a useful supplement in the determination of microbial diversity." Applied And Environmental Microbiology 66(2): 549-557. Wade, P. R., V. N. Burkanov, M. E. Dahlheim, N. A. Friday, L. W. Fritz, T. R. Loughlin, S. A. Mizroch, M. M. Muto, D. W. Rice, L. G. Barrett-Lennard, N. A. Black, A. M. Burdin, J. Calambokidis, S. Cerchio, J. K. B. Ford, J. K. Jacobsen, C. O. Matkin, D. R. Matkin, A. V. Mehta, R. J. Small, J. M. Straley, S. M. McCluskey, G. R. VanBlaricom and P. Clapham (2007). "Killer whales and marine mammal trends in the North Pacific - A re-examination of evidence for sequential megafauna collapse and the prey-switching hypothesis." Marine Mammal Science 23(4): 766-802. Wang, S. W., S. J. Iverson, A. M. Springer and S. A. Hatch (2009). "Spatial and temporal diet segregation in northern fulmars (Fulmarus glacialis) breeding in Alaska: insights from fatty acid signatures." Marine Ecology Progress Series 377: 299-307. Watson, J. D. (1972). "Origin of concatameric T7 DNA." Nature New Biology 239: 197-201. Wells, R. S., H. L. Rhinehart, L. J. Hansen, J. C. Sweeny, F. I. Townsend, R. Stone, D. R. Casper, M. D. Scott, A. A. Hohn and T. K. Rowles (2004). "Bottlenose Dolphins as Marine Ecosystem Sentinels: Developing a Health Monitoring System." EcoHealth 1: 246-254. Western, D. (1979). "Size, life history and ecology in mammals." African Journal of Ecology 147: 185-204. Will, K. W. and D. Rubinoff (2004). "Myth of the molecule: DNA barcoding for species cannot replace morphology for identifications and classification." Cladistics 20: 47-55.

148

Woischnick, M. and C. T. Moraes (2002). "Pattern of Organization of Human Mitochondrial Pseudogenes in the Nuclear Genome." Genome Research 12(885-893). Wright, B. E. (2010). "Use of chi-square tests to analyse scat-derived diet composition data." Marine Mammal Science 26(2): 395-401. Ying, T. C. (2005) Evaluation of telomere length as an age marker in marine teleosts. MSc Thesis. University of Hong Kong. Hong Kong. Zaidi, R. H., Z. Jaal, N. J. Hawkes, J. Hemingway and W. O. C. Symondson (1999). "Can multiple-copy sequences of prey DNA be detected amongst the gut contents of invertebrate predators?" Molecular Ecology 8: 2081-2087. Zhang, W., C. Yi, W. Bao, B. Lui, J. Cui, H. Yu, X. Cao, M. Gu, M. Lui and Z. Cheng (2005). "The Transcribed 165-bp CentO Satellite Is the Major Functional Centromeric Element in the Wild Rice Species Oryza punctata." Plant Physiology 139: 306-315. Zischler, H. (2000). "Nulcear integrations of mitochondrial DNA in primates: Inferences of associated mutational events." Electrophoresis 21: 531-536. Zug, G. R., A. H. Wynn and C. Ruckdeschel (1986). Age Determination of Loggerhead Sea Turtles, Caretta caretta, by Incremental Growth Marks in the Skeleton.Smithsonian Contributions to Zoology. Number 427.Smithsonian Institution Press.City of Washington. Zuur, A. F., E. N. Ieno, N. J. Walker, A. A. Saveliev and G. M. Smith (2009). Mixed Effects Models and Extensions in Ecology with R.Statistics for Biology and Helth.M. Gail, K. Krickeberg, J. M. Samet, A. Tsiatis and W. Wong Springer.New York.