If humans are 99.9 percent genetically identical, as President Bill Clinton is fond of asserting when he extols the Human Genome Project, that 10th-of-a-percent difference has a lot of explaining to do. How does genetic variation determine a person's unique physical traits? Can it predict someone's susceptibility to a disease?
Such questions, pertaining to the present or future, are what occupy most human geneticists. A small group, however, studies genetic variation as a clue to the past. Sometimes called molecular anthropologists, these researchers use DNA polymorphisms, or markers, to hypothesize about human evolution and population migrations. An estimated 20 labs worldwide are intensively engaged in this work, and a growing number are peripherally involved.
Two developments lately have boosted their endeavors. Scientists have discovered enough useful DNA markers over the past 20 years to form a critical mass. And techniques and equipment, some of which are spillover from genome sequencing projects, have become increasingly efficient.

Freighted with scientific jargon and numbers, papers in molecular anthropology nevertheless afford fascinating glimpses into history1--"National Geographic magazine kind of stuff," as one researcher puts it. Some findings are also of biomedical use, helping target the genetic causes of diseases. This justifies limited public funding in a field that often depends on private grants.
The interests of molecular anthropologists range widely, from small tribes to humankind, from the emergence of Homo sapiens to population dislocations in recent centuries. The work of Luigi Luca Cavalli-Sforza, a professor emeritus of genetics at Stanford University School of Medicine, typifies the ambitions of the field. For the past two decades, he has been examining genetic variation to deduce migrations into Europe. Among his most striking conclusions is that a vast influx of farmers from the Middle East occurred during the Neolithic period.2
For Cavalli-Sforza, a founder of molecular anthropology in the 1950s when protein (rather than DNA) polymorphisms were the object of attention, says his work is far from complete: "We have really just scratched the surface. We need more markers. We need to study more individuals than we have done. We need to have better collections [of samples]." Charting variation in many genes is also crucial, according to Aravinda Chakravarti, a genetics professor at Case Western Reserve University School of Medicine. "In a process of evolutionary events, not every gene behaves in exactly the same way," he observes.
Genome Diversity Projects
In 1991, Cavalli-Sforza cofounded the Human Genome Diversity Project (HGDP) to coordinate the gathering and analysis of DNA samples from ethnic groups around the world. But HGDP soon became embroiled in ethical and political controversies, and a National Research Council report in 1997, while largely favorable to the project,3 didn't give it a needed boost. Large-scale government funding for a worldwide HGDP hasn't come through.
"The present trend is for countries to take their own initiatives," says Cavalli-Sforza. He describes North America, Europe, Japan, Russia, and most African nations as open to such research. He singles out HGDP programs in China, Israel, and Pakistan as excellent. But he adds, "A number of indigenous groups, including many American Indians, have declared their unwillingness to collaborate." And India, the world's second most populous nation, won't allow its citizens' DNA to be sent abroad.
Cavalli-Sforza is enthusiastic about a new project by the Paris-based Centre d'Etude du Polymorphisme Humain (CEPH). "My idea was for all the labs which, like mine, have generated cell lines in various parts of the world to donate these cell lines to CEPH," he says. "And CEPH can grow them, make DNA, and distribute it to research workers." Distribution is slated to begin this fall after CEPH has collected 1,000 cell lines representing about 50 populations. The DNA will be available at modest prices that are expected to finance this nonprofit project. Researchers will be asked to enter the results of their studies of the DNA into CEPH's database.
Y Chromosome
Cavalli-Sforza's research has concentrated on the Y chromosome in the past decade. Because only 1 percent of the chromosome, at its tip, recombines with the X chromosome during meiosis, Y polymorphisms are an ideal source of information on men's paternal ancestry. Until several years ago, few Y markers were known. The problem was twofold: Mutations may be rarer on the Y chromosome than on autosomes, and detection of markers via nonautomated DNA sequencing and some gel-based techniques was slow and laborious. "We couldn't live on finding one marker a year," recalls Peter A. Underhill, now a senior research scientist in Cavalli-Sforza's lab. "It was so hard to motivate people to come in every day, do experiments, and--nothing, nothing, nothing." Cavalli-Sforza's research has concentrated on the Y chromosome in the past decade. Because only 1 percent of the chromosome, at its tip, recombines with the X chromosome during meiosis, Y polymorphisms are an ideal source of information on men's paternal ancestry. Until several years ago, few Y markers were known. The problem was twofold: Mutations may be rarer on the Y chromosome than on autosomes, and detection of markers via nonautomated DNA sequencing and some gel-based techniques was slow and laborious. "We couldn't live on finding one marker a year," recalls Peter A. Underhill, now a senior research scientist in Cavalli-Sforza's lab. "It was so hard to motivate people to come in every day, do experiments, and--nothing, nothing, nothing."
The situation brightened considerably in 1995 when Underhill and Stanford biochemist Peter J. Oefner developed denaturing high-performance liquid chromatography (DHPLC). This method detects polymorphism-containing DNA heteroduplexes when they bind differently to columns than homoduplexes do. Screening samples with DHPLC, Underhill and Oefner have discovered many Y markers,4 which they're using to reexamine Cavalli-Sforza's theories about migrations into Europe. Less than half a percent of the 59-megabase Y chromosome has been scanned, but 200 single nucleotide polymorphisms (SNPs), 40 microsatellites, and two minisatellites have already been detected, according to Michael F. Hammer, an associate professor of anthropology at the University of Arizona in Tucson. (Satellites are tandemly repeated DNA sequences that may display great person-to-person variability in the number of repeats.)
Though Y-marker analysis is new, the New York Times recently trumpeted that scientists could trace human lineages back to "10 sons of a genetic Adam."5 Hammer is more cautious, noting that the idea that all men descend from 10 individuals, as if no other men lived at that time, "is a statistical artifact of the way the data are analyzed." Indeed, molecular anthropologists readily acknowledge how speculative their hypotheses are. Models may rest on shaky assumptions about DNA mutation rates and past human reproductive behavior.
Mitochondrial DNA
Tracing maternal ancestry through mitochondrial DNA (mtDNA) dates back to the 1970s, a decade before Y-marker analysis began in earnest. Human eggs hold 100,000 mitochondria, each with a circular mtDNA of 16,569 nucleotides. This DNA is passed exclusively from mother to offspring. How does a mutation in one mtDNA of one individual replace all the nonmutant mtDNA in a population's mitochondria? In a process called replicative segregation, dividing cells bequeath more mutant mtDNAs to some daughter cells. Countless mitoses magnify this effect until the nonmutant DNA is eliminated.
Thousands of mtDNA polymorphisms have been found, many in the organelle's hypervariable control region, according to Douglas C. Wallace, director of the center for Molecular Medicine at Emory University School of Medicine. The Center maintains a searchable database that listed about 1,400 mtDNA markers as of the end of June.
Analysis of such polymorphisms led to the 1987 proposal by Allan C. Wilson at the University of California, Berkeley, that an "Eve"--a source of all people's mtDNA--lived 200,000 years ago.6 Since that proposal, Wallace and his colleagues have grouped humans into 27 lineages based on mtDNA markers.7
The chronology of the branching of these lineages is still a point of contention between some paleontologists and molecular anthropologists. But the debate isn't totally polarized. "There are excellent physical anthropologists who feel that the fossil data are quite consistent with the molecular data," Wallace notes.
Two years ago, his team generated a buzz when it found that some American Indians display a pattern of mtDNA markers consistent with a European origin. That result has been replicated, he says, and assiduous attempts to detect the same pattern of markers in Asia--the presumed origin of native Americans--have been unsuccessful.
Wallace's team is now on the cusp of finishing a 13-year project to completely sequence about 50 mtDNAs from all the major human lineages. "That's allowing us to look at very detailed questions, like the radiation of individual small populations, and very close time-frame questions," he says. He hopes to submit the results of the project for publication this summer.

"People are just too hung up on mitochondrial DNA and Y chromosomes," complains Kenneth K. Kidd, a genetics professor at Yale University School of Medicine. "Everybody thinks recombination is a problem, diploidy is a problem. And they're not. They're easily overcome. They just require more sophisticated analysis."
Kidd confesses to having mixed feelings about what he perceives as this neglect of autosomes. It means less competition for his work on autosomal genes in humans and chimps. Yet more autosomal studies are crucial for advancing the field of molecular anthropology: The 22 autosomes, after all, harbor the lion's share of polymorphisms.
"Genes on the mitochondrial genome or the Y chromosome don't unambiguously allow you to infer population history," notes Andrew G. Clark, a biology professor at Pennsylvania State University. "That's because there's a lot of stochasticity, a lot of chance, that goes on in sampling of those genomes from generation to generation. What the autosomal genes get us is many more realizations of genes passing through history. If we look at enough of them, we'll be able to get a good call on the true population history." Especially ripe for examination, Clark adds, are autosomal regions with low rates of recombination, which are just now being identified.
Kidd's lab maintains a rapidly growing database containing gene frequency data on more than 175 markers and 60 populations. The database, named ALFRED, can be accessed at the Web site info.med.yale.edu/genetics/kkid
Unlike ALFRED, the SNP Consortium's database of more than 100,000 mapped markers at snp.cshl.org is not broken down by ethnic group and hence is less useful for molecular anthropologists. Last month, however, the consortium began comparing the frequencies of 120,000 SNPs in Africans, Asians, Caucasians, and pre-Columbians, according to chairman Arthur Holden.
In a series of studies building on a seminal 1996 paper,8 Kidd and his colleagues discovered that chromosomes in African populations show more variation and randomness than chromosomes in non-African groups. Kidd explains this finding as follows: In Africa, genes have had time to be scrambled by recombination. here, the genetic patterns inherited from the few founders of a population haven't had time to be scrambled. Thus, humans must have originated in Africa, where populations are oldest. Small migrating groups then populated the rest of the world. Kidd, who investigates alcoholism and psychiatric disorders, also argues for the medical value of characterizing populations, as molecular anthropologists do in the course of their work. In seeking the genetic origins of diseases, "you can hardly interpret the data in a mixed population," he warns. He regrets that studies routinely use such populations.9
Biomedical researchers must also know more about population than its degree of homogeneity. In associational gene cloning, Kidd notes, "If you're studying an African population, you have to have five or six times more markers than if you're studying a European population." The reason is that African chromosomes have been more scrambled by recombination. As a result, genetic markers are likely to be associated with (i.e., found in a population disproportionately together with) a putative disease-promoting polymorphism only if they're very close to that polymorphism. A high density of markers is needed to ensure such proximity.
Molecular anthropology oscillates between investigations of "macro" topics, such as evolution and continental migration, and "micro" topics, such as whether two populations are genetically related. For the near future, Cavalli-Sforza foresees a shift toward regional work "because you can do a better job. But eventually you have to be able to connect all the regions." Molecular anthropology, meanwhile, is a "very, very hot" research area, according to Wallace, who notes that it has prompted three Cold Spring Harbor symposia in the last five years and many international meetings. The public is also starting to become aware of the field. "I get a lot of E-mail from regular folks out there, nonscientists, who are interested in this kind of work," Hammer says. "I think people have a growing interest in genealogy, a growing interest in who they are genetically and whom they're related to."
Satisfying that curiosity could stir up ethnocentrism and bigotry, but the scientists interviewed for this article recalled few such cases. "The genetics is telling a story that does not support racist attitudes," contends Hammer.
Nevertheless, Underhill remembers occasionally receiving hate mail when a study is publicized in the press. Tell some people they're "related to Africans, and then you get the neo-Nazi kooks," he observes. Though he doesn't set out to threaten people's "origin stories," Underhill maintains, "As a scientist, I follow the data." S

• K. Owens, M.-C. King, "Genomic views of human history," Science, 286:451-3, 1999.
• L.L. Cavalli-Sforza, Genes, Peoples, and Languages (New York, North Point Press, 2000). For an approach to this topic using mitochondrial DNA, see L. Simoni et al., "Geographic patterns of mtDNA diversity in Europe," American Journal of Human Genetics, 66:262-78, 2000.
• D. Steinberg, "NIH jumps into genetic variation studies," The Scientist, 12[2]:1, Jan. 19, 1998
  
  
• See, e.g., P.A. Underhill et al., "Detection of numerous Y chromosome biallelic polymorphisms by denaturing high-performance liquid chromatography," Genome Research, 7:996-1005, 1997.
• N. Wade, "The human family tree: 10 Adams and 18 Eves," New York Times, May 2, 2000, p. F1. 6. R.L. Cann et al., "Mitochondrial DNA and human evolution," Nature, 325:31-6, 1987.
• D.C. Wallace, "Mitochondrial DNA variation in human evolution and disease," Gene, 238:211-30, 1999.
• S.A. Tishkoff et al., "Global patterns of linkage disequilibrium at the CD4 locus and modern human origins," Science, 271:1380-7, 1996.
• See, e.g., A.M. Kang et al., "Global variation of a 40-bp VNTR in the 3' untranslated region of the dopamine transporter gene (SLC6A3)," Biological Psychiatry, 46:151-60, 1999.