protein functions prediction

Swiss Institute of BioinformaticsInstitut Suisse de Bioinformatique

LF-2002.10

Protein functions prediction


LF-2002.10

Introduction Signal peptides Transmembrane

regions and topology PTM (post-

translational modifications)

Low complexity and biased regions

Repeats Coils

Secondary structure Antigenic peptides Domain/Motifs Tools The EMBOSS package


LF-2002.10

Different techniques Algorithms

Sliding window, Nearest Neighbor Patterns, regular expression Weight matrices HMM, profiles Neural Networks Rules


LF-2002.10

Sliding windowTHISISATESTSEQVENCETHATDISPLAYSTHESLIDINGWINDQW

Score1Score2Scoren

Width or Size=11, Step=5Results are usually displayed as a graph, see example ->


LF-2002.10

Patterns / regular expression Pattern: <A-x-[ST](2)-x(0,1)-{V} Regexp: ^A.[ST]{2}.?[^V] Text: The sequence must start with an

alanine, followed by any amino acid, followed by a serine or a threonine, two times, followed by any amino acid or nothing, followed by any amino acid except a valine.

Simply the syntax differ…


LF-2002.10

Weight matrices (PSSM)


LF-2002.10

HMM / profiles


LF-2002.10

Neural NetworksGeneral principle: Example:


LF-2002.10

Signals found in proteins N-ter

exportation - secretion mitochondria chloroplast

internal NLS (nuclear

localization signal)

C-ter GPI-anchor (Glycosyl

Phosphatidyl Inositol) other membrane

anchors (see PTM) other unknown ?


LF-2002.10

Signals detection tools SignalP MitoProt ChloroP Predotar PSort TargetP Sigcleave (EMBOSS)

Big-PI DGPI


LF-2002.10

Transmembrane regions Detection (signal peptide, hydropathy, helices) Organisation (topology)


LF-2002.10

Transmembrane detection tools

TMHMM TMPred TopPred2 DAS HMMTop Tmap (EMBOSS)


LF-2002.10

Post translational modifications

Phosphorylation S - T - Y

N-glycosylation N

O-glycosylation S - T - (HO)K

Acetylation, methylation D - E - K

Sulfation Y

Farnesylation, myristylation, palmitoylation, geranylgeranylation, GPI-anchor C - Nter - Cter

Ubiquitination and family K - Nter

Inteins (protein splicing) Pre-translational

Selenoprotein C


LF-2002.10

PTM detection Pattern prediction

(PROSITE) Short or weak signal Frequent hit producer Best method is

experimental MS/MS detection

Most method use « rules » joining pattern detection and knowledge to predict sites.

NetOGlyc - Prediction of type O-glycosylation sites in mammalian proteins

DictyOGlyc - Prediction of GlcNAc O-glycosylation sites in Dictyostelium

YinOYang - O-beta-GlcNAc attachment sites in eukaryotic protein sequences

NetPhos - Prediction of Ser, Thr and Tyr phosphorylation sites in eukaryotic proteins

NMT - Prediction of N-terminal N-myristoylation

Sulfinator - Prediction of tyrosine sulfation sites


LF-2002.10

Low complexity regions repeats compositional bias PEST


LF-2002.10

Low complexity / Repeats DUST (DNA) / SEG

de novo detection RepeatMasker (DNA)

search collection REP

search collection REPRO, Radar

de novo detection PEST, PESTFind

de novo detection

EMBOSS (DNA) einverted equicktandem etandem palindrome

EMBOSS (protein) oddcomp


LF-2002.10

Coils Helix of helix

coiled-coil Leu-zipper


LF-2002.10

Coils detection COILS

Weight matrices Paircoil, Multicoil

Pairwise correlation Marcoil

HMM Pepcoil (EMBOSS)

Weight matrices


LF-2002.10

Secondary structure Structure to

predict Alpha-helices Beta-sheets Turns Random coil

Garnier (EMBOSS) PHD DSC PREDATOR NNSSP Jpred Jnet Many others


LF-2002.10

Antigenic peptide Peptides binding to MHC class I

8, 9, 10 mers class II

15 mers (3+9+3) Depend highly on MHC

type

Use of experimental knowledge Databases of known

peptides

SYFPEITHI HLA_Bind (BIMAS) MAPPP combined expert Antigenic (EMBOSS) Many more

Prediction of proteasome cleavage sites

NetChop PaProc


LF-2002.10

Domain / Motif All the protein domain

descriptors PROSITE PFAM SMART PRODOM BLOCKS PRINTS …

Federation: InterPro Many techniques

Patterns, Regexp PSSM (PSI-BLAST) Profiles HMM


LF-2002.10

Other Tools You can find some of them on our servers

www.ch.embnet.org Or on ExPASy server

www.expasy.org/tools Or ask Google!!

www.google.com


LF-2002.10

European Molecular Biology Open Software Suite


LF-2002.10

Free Open Source (for most Unix plateforms) GCG successor (compatible with GCG file format) More than 100 programs Easy to install locally

but no interface, requires local databases Unix command-line only

Interfaces Jemboss, www2gcg, w2h, wemboss … (with account) Pise, EMBOSS-GUI (no account)

Access: http://www.emboss.org

http://www.emboss.org/

http://www.emboss.org/


LF-2002.10

http://www.hgmp.mrc.ac.uk/Software/EMBOSS/Jemboss/index.html

http://www.hgmp.mrc.ac.uk/Software/EMBOSS/Jemboss/index.html


LF-2002.10

Pise a tool to generate Web interfaces for Molecular Biology programs

http://emboss.ch.embnet.org:8080/Pise

http://emboss.ch.embnet.org:8080/Pise


LF-2002.10

http://bioinfo.pbi.nrc.ca:8090/EMBOSS/

GUI (Canada)

http://bioinfo.pbi.nrc.ca:8090/EMBOSS/index.html


LF-2002.10

Format USA 'asis' :: Sequence [start : end : reverse] Format :: '@' ListFile [start : end : reverse] Format :: 'list' : ListFile [start : end : reverse] Format :: Database : Entry [start : end : reverse] Format :: Database - SearchField : Word [start : end : reverse] Format :: File : Entry [start : end : reverse] Format :: File : SearchField : Word [start : end : reverse] Format :: Program Program-parameters '|' [start : end : reverse]

Example: fasta::Swissprot:UBP5_HUMAN[200:300]

Databases Any can be added, use showdb to display the available databases

Some details


LF-2002.10

Some tools for DNA redata Search REBASE for enzyme name, references, suppliers etc remap Display a sequence with restriction cut sites, translation etc restover Finds restriction enzymes that produce a specific overhang restrict Finds restriction enzyme cleavage sites showseq Display a sequence with features, translation etc silent Silent mutation restriction enzyme scan cirdna Draws circular maps of DNA constructs lindna Draws linear maps of DNA constructs revseq Reverse and complement a sequence …


LF-2002.10

Example: remapECLAC E.coli lactose operon with lacI, lacZ, lacY and lacA genes. Hin6I TaqI | HhaI | Bsc4I | Bsu6I | | Hin6I | BssKI | | | HhaI AciI | | BsiSI \ \ \ \ \ \ \ \ GACACCATCGAATGGCGCAAAACCTTTCGCGGTATGGCATGATAGCGCCCGGAAGAGAGT 10 20 30 40 50 60 ----:----|----:----|----:----|----:----|----:----|----:----| CTGTGGTAGCTTACCGCGTTTTGGAAAGCGCCATACCGTACTATCGCGGGCCTTCTCTCA / / / / / / / /// | TaqI | Hin6I AciI | | ||BssKI Bsc4I HhaI | | |BsiSI | | Bsu6I | Hin6I HhaI# Enzymes that cut Frequency Isoschizomers AciI 1 Bsc4I 1 BsiSI 1 BssKI 1 Bsu6I 1 HhaI 2 Hin6I 2 HinP1I,HspAI TaqI 1# Enzymes that do not cutAclI BamHI BceAI Bse1I BshI ClaI EcoRI EcoRII Hin4I HindII HindIII HpyCH4IV KpnI NotI


LF-2002.10

Example: cirdna File: ../../data/data.cirpStart 1001End 4270grouplabelBlock 1011 1362 3ex1endlabellabelTick 1610 8EcoR1endlabellabelBlock 1647 1815 1endlabellabelTick 2459 8BamH1endlabellabelBlock 4139 4258 3ex2endlabelendgroupgrouplabelRange 2541 2812 [ ] 5AluendlabellabelRange 3322 3497 > < 5MER13endlabelendgroup


LF-2002.10

Exercises DEA Exercises web based sequence analysis The goal of this exercise is to use web based tools for protein sequence analysis

a) Take this TrEMBL sequence (Q9X252) and try a BLAST against swissprot with the complete protein or with the first 70 residues. Explain the difference. Use TMPred, SignalP, and COILS to help you.

b) Pass this sequence through PFSCAN and search all databases. Compare with this command on ludwig-sun1/2: hits -b "prf pat pfam" tr:Q9X252

c) use the different profile, motifs, pattern databases to get more information about the domain(s) you found.

d) How do you evaluate the PRINTS tropomyosin annotation in this TrEMBL entry (Q9WZH0)? List of useful links:

basic BLAST or advanced BLAST or PSI-BLAST TMPred prediction tool for transmembrane regions (or TMHMM) COILS prediction tool for coiled-coil regions SignalP prediction tool for signal-peptide cleavage site

Profile, domain, motifs databases and search sites: PFSCAN InterPro (Pfam, PRINTS, PROSITE, SMART) HITS

protein functions prediction

Documents