1499

tivity, especially the latter, have revealedthat individuals both compete and cooper-ate by making inferences about what othersknow and intend (10, 11). These studieshave revolutionized our understanding ofwhat chimpanzees think and feel, raisingprofound philosophical questions about thenature of thought without language, as wellas ethical questions concerning the rightsand welfare of these animals (12).

Constraining our continued understand-ing of this wonderful animal is one annoyinghurdle: our own species. In the very nearfuture, we may ironically face the possibilityof having a detailed map of the chimpanzeegenome, but no individuals to study. Illegalhunting, the bushmeat trade, and deforesta-

tion are destroying chimpanzee populations(see, for example, www.chimpcollabora-tory.org). If the same amount of effort that isgoing into genetic analyses went into chim-panzee conservation and behavioral biology,not only would we save this species fromextinction, but we would write the mostdetailed story of our past—as rich as theBible, but grounded in science.

References and Notes

1. The Chimpanzee Sequencing and AnalysisConsortium, Nature 437, 69 (2005).

2. J. Goodall, The Chimpanzees of Gombe: Patterns ofBehavior (Belknap/Harvard, Cambridge, MA, 1986).

3. M. A. Huffman, R. W. Wrangham, in ChimpanzeeCultures. R. W. Wrangham, W. C. McGrew, F. B. M. deWaal, P. G. Heltne, Eds. (Harvard Univ. Press,Cambridge, MA, 1994), pp. 129–148.

4. A.Whiten et al., Nature 399, 682 (1999).5. T. Matsuzawa, Primate Origins of Human Cognition and

Behavior (Springer-Verlag, Berlin, 2002).6. M. L. Wilson, M. D. Hauser, R. W. Wrangham, Anim.

Behav. 61, 1203 (2001).7. R. W. Wrangham, D. Peterson, Demonic Males

(Houghton Mifflin, New York, 1996).8. M. D. Hauser, Wild Minds: What Animals Really Think

(Holt, New York, 2000).9. F. B. M. de Waal, Chimpanzee Politics: Power and Sex

Among Apes (Harper & Row, New York, 1982).10. D. Premack, A. Premack, Original Intelligence

(McGraw-Hill, New York, 2002).11. M. Tomasello, J. Call, B. Hare, Trends Cognit. Sci. 7, 153

(2003).12. P. Cavalieri, P. Singer, The Great Ape Project: Equality

Beyond Humanity (St. Martin’s, New York, 1994).13. I thank three of my closest colleagues for comments

on this essay: F. de Waal, B. Hare, and R.Wrangham.

10.1126/science.1111421

When the Human Genome Projectwas established in 1991, the plan-ners wisely included sequencing

the genomes of model organisms in the pro-ject’s goals. At that time, the only nonhu-man mammalian genome scheduled forsequencing was that of the laboratorymouse. Although the relevance of themouse genome for interpreting the humansequence was beyond dispute, some biolo-gists were disappointed that no nonhumanprimate genome had been included. Theremarkable similarity of the chimpanzeegenome to that of humans was already pre-dicted from overall DNA comparisons, andit seemed clear that questions about thegenetic basis for human uniqueness wouldeventually require detailed comparisonswith the genomes of great apes (1), ourclosest evolutionary relatives. A formalpresentation of the need for sequencing thechimpanzee genome was published in 1997(2). Soon thereafter it was pointed out (3)that there should also be a project toincrease our knowledge of the great ape“phenome” (the complete body of informa-tion about an organism’s phenotype undervarious environmental conditions), aboutwhich very little is known. Scientists from avariety of disciplines rallied in support of

sequencing the chimpanzee genome, alsociting biomedical reasons and the potentialimportance for proper care and conserva-tion of great apes (4, 5).

We now have a draft sequence of thecommon chimpanzee genome (Pantroglodytes) and a detailed comparisonwith the human genome (6). The resultsinclude extensive information on compar-ative genomics, such as the number of sin-gle base pair and insertion/deletion differ-ences and transposable elements uniqueto either human or chimpanzee. Thereport clarifies much previously conflict-ing or confusing information in existinghuman nucleotide sequence databanks andaddresses several important questionsabout genomic and population evolutionmechanisms. It also adopts a rationalorthologous chromosomal numbering sys-tem to facilitate comparisons of human andape genomic organization (7).

Can we now provide a DNA-basedanswer to the fascinating and fundamentalquestion, “What makes us human?” Not atall! Comparison of the human and chim-panzee genomes has not yet offered anymajor insights into the genetic elements thatunderlie bipedal locomotion, a big brain, lin-guistic abilities, elaborated abstract thought,or any other unique aspect of the human phe-nome. This state of affairs may seem disap-pointing, but it is merely the latest exampleof a generalization that genomics researchhas already established—interpretation ofDNA sequences requires functional infor-mation from the organism that cannot be

deduced from sequence alone. Functionalgenomics investigations must determinewhere a gene is expressed within an organ-ism, when it is expressed during develop-ment and life history, and what the level ofexpression is at various times. Furthermore,these data must be integrated with informa-tion about the related phenotypes, as well ascritical environmental influences underwhich the genotype generates the phenotype(see the figure).

There are three general reasons for sub-stantially increasing research on chim-panzees (and the other great apes—bono-bos, gorillas, and orangutans): First, tounderstand the contribution of genomicDNA to human and great ape evolution;second, to improve our understanding ofhuman and ape phenomes (at all levels,from molecular to behavioral to states ofdiseases); and third, to help preserve popu-lations of these important human relatives.These goals must be pursued in the face ofchallenging ethical issues that still need tobe resolved by open debate.

Understanding the genetic basis ofuniquely human traits will require increas-ing the accuracy and completeness of thecurrently available chimpanzee genomesequence, as well as sequencing other pri-mate genomes as out-groups. The genomesof the orangutan and the rhesus macaqueare currently being sequenced, but othergenomes are needed to obtain a completepicture. Among other benefits, such multi-species comparisons are essential for iden-tifying human-specific coding and regula-tory regions.

A parallel requirement is the compari-son of human gene expression with thoseof chimpanzees and other primates. Thereare formidable obstacles to achieving thisgoal, the most obvious of which is obtain-ing experimental material from great apes.It is not ethically acceptable to sacrifice agreat ape simply to obtain tissue samples.

G E N O M I C S

Thoughts on the Future

of Great Ape ResearchEdwin H. McConkey and Ajit Varki

E.H. McConkey is in the Department of Molecular,Cellular and Developmental Biology, University ofColorado, Boulder, CO 80309, USA. E-mail:mcconkey@colorado.edu A. Varki is in theDepartments of Medicine and Cellular and MolecularMedicine, University of California, San Diego, La Jolla,CA 92039, USA. E-mail: a1varki@ucsd.edu

www.sciencemag.org SCIENCE VOL 309 2 SEPTEMBER 2005

P E R S P E C T I V E S

Published by AAAS

1500

Studies of great apes should follow guide-lines generally similar to those for researchon human subjects. Thus, there is a needfor new funding to support development ofa network among current holders of cap-tive great apes (including primate facili-ties, great ape sanctuaries, and zoos) toguarantee that tissue samples can beobtained quickly from each great ape thatdies of natural causes, or has to be eutha-nized because of incurablesuffering. Autopsy samplesneed to be preserved for his-tological analysis and used assource materials for studiesof gene expression andcDNA libraries. Such sam-ples can also be used for awide variety of other “omic”comparisons (including pro-teomics, glycomics, andlipomics). Because every tis-sue is made up of multiplecell types, such approachescan still miss important dif-ferences in minor cell types.Thus, parallel histologicalcomparisons must occur,using multiple probes todetect differences.

Examination of adult tis-sues, however, will still notallow us to understand geneexpression and its conse-quences during development, which maywell be the time when many of the crucialdifferences between humans and the greatapes are expressed. This concept was sug-gested decades ago by King and Wilson (8),and studies since then have given us no rea-son to reject that hypothesis. Analysis ofgene expression during prenatal develop-ment can be approached with three experi-mental strategies: transgenesis, stem cells,and direct study of developmental samples.

Transfer of human genes into mice hasbeen fundamental to the analysis of humangene function. Comparative analysis ofhuman and chimpanzee orthologous genesin transgenic mice is now certain to be pur-sued. Of course, there are limits to whatone can deduce about the phenotypiceffects of human or ape genes in mice. Thisis particularly true for the brain, skin,innate immune system, and reproductivesystem, wherein primates have undergoneconsiderable functional divergence fromrodents. However, ethical, fiscal, and prac-tical considerations will make the idea oftransgenic apes moot. Thus, we shouldexpect that the need for transgenic mon-keys will arise, and deciding on the idealmodel for such experiments will not beeasy. Given ethical and practical issues andthe longer generation-time of monkeys,

such studies will require much thought,patience, and long-term funding.

Embryonic stem cell cultures fromhumans are a subject of intense interest.Although creation of stem cell lines fromape embryos will be just as difficult techni-cally, it represents a feasible experimentalapproach that causes no lasting harm to theanimals from which gametes are obtainedfor in vitro fertilization. As technical

progress is made with human stem cells,this knowledge can be applied to chim-panzee stem cells, providing a major sourceof information on gene expression in sev-eral embryonic and differentiated cell types,during various stages of in vitro develop-ment. If current approaches to tissue andorgan engineering with human stem cellsare successful, parallel studies of chim-panzee equivalents could provide furtherresources to study expression of genes andgene products and contribute to treatmentof great ape diseases in the future.

Material for direct analysis of geneexpression during embryonic developmentcan, in principle, be obtained by controlledbreeding and surgical termination of preg-nancies. This approach is already wellestablished for monkeys such as the rhesusmacaque. A thorough study of gene expres-sion during monkey embryonic develop-ment would be expensive, but it should beundertaken (9), perhaps only after studies oftransgenic mice and chimpanzee stem cellshave defined critical experimental ques-tions that cannot be otherwise answered. Wedo not envision this being done with greatapes because of ethical and practical con-siderations. As with humans, however, greatape samples may become available in thecourse of birth control or medical care.

The second reason for expandingresearch on chimpanzees and other apes isthe lack of information on their phenotypes(3). The utility of the human genome hasbeen greatly aided by our vast knowledge ofthe human phenome in areas ranging fromanatomy to cognitive function. In contrast,our knowledge of the great ape phenome isinadequate, except in a few arenas such asbehavior and ecology. Worse, extant infor-mation on great apes is in many scatteredsources spanning the last century, and someaccepted “facts” actually represent folklorederived from misinterpretations or assump-tions made in popular science literature.Thus, there is no easy way to reliably ascer-tain all the known and unknown differencesbetween humans and great apes. One possi-bility (10) is to develop a Web-based“Museum of Comparative Anthropogeny”that would catalog information abouthuman-specific differences from great apesthat is scattered throughout the literature.Having a centralized resource of such infor-mation could lead to new conceptualinsights and multidisciplinary interactionsand also point to ethically acceptable stud-ies that would help to explain human-apedifferences. Regardless, interpreting theresults of functional genomics studies willrequire more information about ape phe-nomes. Thus, there should be substantiallyincreased funding for studies on great apeanatomy, physiology, biochemistry, neuro-biology, cognitive functions, behavior, andecology. All such research should be donefollowing ethical principles like those cur-rently used in human studies. Much can alsobe learned in the course of providing out-standing medical care, as has been the casefor humans. Increased knowledge about apephenomes will likely be helpful for under-standing some human diseases (5).

The third reason for expanding researchon chimpanzees and other great apes is thatthe more we know about these species, thebetter we can care for them. This will be par-ticularly important for captive apes, butcould also have an impact on maintaininghealthy wild populations (for example, byvaccinating them against human diseases). Inthis regard, making practical use of all thefunctional and behavioral knowledge arisingfrom such research will require a significantincrease in financial support for the optimalmaintenance of captive great apes, and tofacilitate survival of currently endangeredwild populations. One way to coordinatefunding for ape research and care is to createa Great Ape Conservation Trust that wouldreceive 10% of all grant funds awarded bygovernment agencies for research on apegenomes, phenomes, or behavior. The Trustcould be administered by an agency that doesnot award research grants; instead, it would C

RED

IT:P

RES

TON

HU

EY/S

CIE

NC

E

2 SEPTEMBER 2005 VOL 309 SCIENCE www.sciencemag.org

Genome

Environment

Phenome

Chimpanzee

Compare

Human

Compare

Selection Selection

What makes us human? This question may be answered by compar-ison of human and chimpanzee genomes and phenomes, and ulti-mately those of other primates. To this end, we need to understandhow genotype generates phenotype, and how this process is influ-enced by the physical, biological, and cultural environment.

P E R S P E C T I V E S

Published by AAAS

1501

CR

EDIT

:PR

ESTO

N H

UEY

/SC

IEN

CE,

AD

APT

ED F

RO

M Z

HA

O E

T A

L. (

9)

award grants only for the support of captiveanimals and the conservation of wild popula-tions. The agency that administers the Trustcould be either a governmental or non-governmental group that already exists, or anew organization with representatives fromvarious interest groups, particularly thosewith firsthand knowledge about conserva-tion issues. The Trust could also be author-ized to solicit and receive funds from non-governmental sources.

With the sequencing of the chimpanzeegenome, we have now reached the end of

the beginning, and can start down the longroad toward fully understanding our rela-tionships to these closest evolutionarycousins. If the road is taken with intellectualand ethical care, there is much to gain, forboth them and us.

References and Notes 1. The term “great ape” is used here in the colloquial

sense. In the commonly used classification, thesespecies are now grouped with humans in the familyHominidae, and humans belong to the tribe Hominini,along with chimpanzees and bonobos.

2. E. H. McConkey, M. Goodman, Trends Genet. 13, 350(1997).

3. A. Varki et al., Science 282, 239 (1998).4. E. H. McConkey,A.Varki, Science 289, 1295 (2000).5. A. Varki, Genome Res. 10, 1065 (2000).6. Chimpanzee Sequencing and Analysis Consortium,

Nature 437, 69 (2005).7. E.H. McConkey, Cytogenet. Genome Res. 105, 157

(2004).8. M.C.King, A. C.Wilson, Science 188, 107 (1975).9. E. H. McConkey, Trends Genet. 18, 446 (2002).

10. M. V. Olson,A.Varki, Science 305, 191 (2004).11. The opinions expressed are strictly those of the

authors.We thank K. Krauter, J. Moore, P. Gagneux, andN. Varki for helpful comments. Supported by the G.Harold and Leila Y. Mathers Charitable Foundation.

10.1126/science.1113863

The size of magneticobjects that can bemanipulated in con-

densed-matter environmentshas decreased over the last50 years, from bulk ferro-magnets to thin f ilms,nanocrystals, clusters, andnow to single atoms andmolecules (1–7). The single-atom or single-moleculeregime is especially interest-ing because magnetismarises in this case from veryfew unpaired electronicspins and is thus quantummechanical in nature. Thisproperty opens new oppor-tunities that range frombasic quantum impuritystudies to quantum informa-tion and spintronics applica-tions (8). Molecular sys-tems, in particular, provide auseful means for “packag-ing” quantum spin centers, because mole-cules are structurally and electronicallyvery flexible (7). This is readily seen in thework of Zhao et al. (9) on page 1542 of thisissue. They show that it is possible to tunethe spin behavior of a magnetic cobalt iontrapped within a single molecule by prun-ing the ligands of the molecule with the tipof a scanning tunneling microscope (STM).

Zhao et al. observed this behavior by

tunneling electrons from the tip of an STMinto a single cobalt phthalocyanine (CoPc)molecule sitting on a gold surface, therebyperforming a type of local electron spec-troscopy. For pristine CoPc molecules theyobserved the d-orbital of the inner cobalt ionto be an energetically broad resonance lyingbelow the Fermi energy (EF, the energy ofthe highest occupied electronic level). Afterplucking hydrogen atoms from the periph-ery of the molecule with their STM, how-ever, the broad d-resonance was replaced bya much narrower resonance pinned at EF,indicating a change in the magnetic natureof the molecule (see the figure).

Monitoring the d-orbital of a magnetic

nanostructure in this way allows one tostudy its magnetic properties. This isbecause magnetism in transition-metal ions(like the cobalt ion in CoPc) arises fromunpaired spins residing in d-orbitals. Whena single cobalt atom or cobalt-carrying mol-ecule contacts a metal surface, the d-orbitalhybridizes with the continuum states of the

surface and broadens energet-ically into a resonance. If theresonance shifts below EF,then electrons are transferredto the d-level, whereas if theresonance shifts above EF,then charge is pulled out of it.The magnetic moment of theion depends on how the d-orbital is f illed with elec-trons, but the precise filling isdifficult to determine whenonly a limited energy range isexperimentally accessible.

This is where a subtle phe-nomenon known as theKondo effect proves useful.The Kondo effect (10) de-scribes the process by whichelectrons from a surroundingsubstrate magnetically screenthe spin of a magnetic ion.This effect is driven by aninteraction between the local-ized spin of the ion and theitinerant spins of the sub-

strate, and induces the magnetic ion toeffectively “capture” an electron spin fromthe substrate and loosely bind it in a net-zero-spin configuration (that is, the twospins cancel). The signature of the newbound state is a narrow resonance (theKondo resonance) that appears at EF andwhose width (the Kondo temperature) givesa measure of the captured spin’s bindingenergy (10). Historically, this effect hasbeen observed in bulk materials containingmagnetic impurities because a change inthe density of states at EF can significantlyaffect bulk properties such as magnetiza-tion, specific heat, and conductivity. Morerecently, the Kondo effect has been

P H Y S I C S

Manipulating Magnetism

in a Single MoleculeMichael F. Crommie

STM tip

Molecule flat on surface

Removed hydrogen atoms

Molecule now standing on surface

Emitted electrons

Kondo screening cloud

Molecular magnetic surgery. (Top) An STM tip is used to snip hydrogen atomsfrom a single cobalt phthalocyanine molecule lying on a gold surface. (Bottom)The trimmed molecule protrudes from the surface and is surrounded by a cloud ofelectrons that represent the Kondo screening cloud about the cobalt ion spin.

The author is in the Physics Department, University ofCalifornia, Berkeley, CA 94720, USA, and the MaterialsSciences Division, Lawrence Berkeley NationalLaboratory, Berkeley, CA 94720, USA. E-mail: crom-mie@berkeley.edu

www.sciencemag.org SCIENCE VOL 309 2 SEPTEMBER 2005

P E R S P E C T I V E S

Published by AAAS