FINGERPRINTING was invented in Bengal, during the Victorian Raj, by British army officers who wanted a sure fire way to tell their colonial subjects apart. Excited by the discovery that fingerprints were a reliable way of identifying individuals, the father of modern eugenics, Francis Galton, hoped to prove they would also reveal an individual鈥檚 鈥渞ace鈥. He and his colleagues pored over thousands of fingerprints taken from people variously categorised as English, Welsh, Jews, Negroes and Basques. By 1892, however, he had to admit defeat: there were no systematic differences in fingerprints among any of the groups.
A century on, researchers have resumed the hunt. This time, however, the fingerprints they are poring over are genetic 鈥 and this time they might not admit defeat so readily. Emboldened by the impact DNA fingerprints have made already on forensic science, geneticists are looking for 鈥渟ignatures鈥 in human DNA that could help forensic teams to make a good guess at a suspect鈥檚 skin colour and facial features. At the moment, DNA tests come into play only after suspects have been arrested, as an aid to establishing their guilt or innocence. But if the forensic geneticists looking for the new DNA signatures have their way, forensic teams could soon be predicting the 鈥渞acial鈥 origins of suspects well before anyone is arrested.
That, needless to say, will be about as welcome in many quarters as Galton鈥檚 ideas about race and eugenics. The question critics will ask is not just how accurate such tests would be, but whether the whole notion makes any sense scientifically when the little human genetic diversity that exists is fairly evenly distributed throughout the species.
Advertisement
There are other, political worries too. In an age of 鈥渆thnic cleansing鈥 and resurgent nationalism, wouldn鈥檛 any kind of DNA test for 鈥渞ace鈥 鈥 even one based on the tiny fraction of human genes that do vary with skin colour and other visible 鈥渞acial鈥 or 鈥渆thnic鈥 characteristics 鈥 create more problems than it solved?
Naturally enough, proponents talk up the social benefits of developing such tests. Imagine, for instance, if forensic scientists could issue 鈥淧hotofits鈥 of suspects based on samples of blood, semen or hair taken from the scenes of crimes. It鈥檚 not as farfetched as it sounds. Such Photofits would require detailed information about the genes that influence the visible markers of 鈥渞ace鈥 鈥 hair colour, eye colour, skin colour, facial features and stature. And in time, some of that information is sure to emerge from international efforts to map and sequence the human genome.
Only skin deep
Take human skin colour, for example, which varies in a continuum through every possible shade. It seems to be influenced by three or four gene pairs that work together, in an incremental or 鈥渁dditive鈥 fashion. Even so, genetic differences probably account for only some 70 or 80 per cent of the variation in colour, says Ashley Robins of the University of Cape Town in South Africa and author of Biological Perspectives on Human Pigmentation. 鈥淪kin colour is a complex phenomenon,鈥 he says. We usually think of it as entirely determined by genetics, but at least 20 per cent of the variation is caused by environmental factors such as nutrition, hormones and lifelong exposure to sunlight. The same is true, he says, of eye and hair colour.
So even with some of the relevant gene sequences to hand, creating Photofits of suspects would be far from trivial. Most human physical characteristics depend not on one but several genes, and working out which genes influence which characteristic (and by how much) would require what Alec Jeffreys, a geneticist at the University of Leicester, who invented DNA fingerprinting in 1985, describes as a 鈥減rofound revolution鈥 in molecular genetics. 鈥淚ndeed,鈥 he says, 鈥渟uch DNA analyses may eventually prove in practice to be impossible.
But cruder tests that give probabilistic predictions are already going ahead. These are based on stretches of DNA that have nothing to do with skin colour or facial features. Walter Bodmer, director of the Imperial Cancer Research Fund in London, predicts that within a few years, forensic scientists analysing a sample taken from the scene of a crime will be able to say that 鈥渢here is a fifty times higher than average chance that the person is a Bengali, or a Welshman, or whatever鈥. Indeed, the prototypes of such DNA tests for 鈥渞ace鈥 already exist.
In 1993, a British forensic scientist published what is perhaps the first DNA test explicitly acknowledged to provide 鈥渋ntelligence information鈥 along 鈥渆thnic鈥 lines for 鈥渋nvestigators of unsolved crimes鈥. Ian Evett, now at the Home Office鈥檚 forensic science laboratory in Birmingham, and his colleagues in the Metropolitan Police, claim that their DNA test can distinguish between 鈥淐aucasians鈥 and 鈥淎fro-Caribbeans鈥 in 85 per cent of cases. Evett鈥檚 test, published in the Journal of the Forensic Science Society, draws on apparent genetic differences in three sections of human DNA (This Week, 23 January 1993). Like most stretches of human DNA used for forensic typing, each of these three regions differs widely from person to person irrespective of race. But by looking at all three, say the researchers, it is possible to estimate the probability that someone belongs to a particular racial group.
At best, however, this type of test will only ever be able to produce rough predictions about 鈥淏engalis鈥 versus 鈥淲elshmen鈥. The DNA signatures they depend on are never found exclusively in one particular racial group. What does vary from group to group is the frequency with which such signatures occur. And even then there are no sharp divisions between racial groups that could be interpreted as marking DNA boundaries 鈥 just gradual and continuous shifts in the frequencies of certain signatures. Analyses of proteins found in blood tell the same story. Neither blood proteins nor the genes that encode those proteins define boundaries between 鈥渞acial鈥 groups. Instead, all these molecular markers reveal are continuous gradations in gene frequencies.
Just an illusion
In short, the notion of homogeneous, well-defined racial types is illusory, the product of cultural categorisations allied with small packages of mostly unidentified genes that evolved to adapt skin, hair and facial features to local climatic conditions. Such genes, according to Jonathan Kingdon, an evolutionary biologist at the University of Oxford, 鈥済ive superficial similarity to people who are actually of very diverse origins鈥. All this, however, may readily be lost from view if DNA tests designed to give probabilistic predictions of physical appearance bolster popular belief that it is scientifically sound to test for racial types. As a result of molecular biology 鈥渞ace will regain a new legitimacy鈥, fears Paul Rabinow, an anthropologist at the University of California at Berkeley. Genetic data will seem more real, more definitive, than the old cultural classifications.
To develop even crude DNA tests for 鈥渞ace鈥, geneticists need samples of DNA from many individuals from many different populations. But law enforcers are eager to help. Earlier this year, Britain鈥檚 police force set up the world鈥檚 first national DNA database. In the crime-torn US, law enforcement agencies in more than 20 states have established forensic DNA databanks. And a nationwide computerised network called CODIS, or Combined DNA Identification System, is now coming online. Designed to link regional DNA forensic centres, CODIS is being set up by the FBI.
Body of evidence
Another rich source of material for what is sometimes called 鈥渇orensic anthropology鈥 is the human genome project 鈥 particularly its recently launched offshoot, known as the Human Genome Diversity Project. 杏吧原创s working on this international collaboration intend to collect DNA samples from more than 400 isolated ethnic groups scattered throughout the world (see 鈥淕enes from a disappearing world鈥, New 杏吧原创, 29 May 1993). They will focus on ones likely to be the most distinctive in genetic terms: isolated populations with distinct cultures and languages which are under threat of extinction. Already, cell lines containing the genetic material of the Baika pygmies of Central Africa and the hill people of New Guinea thrive in a laboratory at Yale University.
In analysing such genetic material, researchers will focus on DNA regions known to vary the most from individual to individual. The DNA of two genetically unrelated individuals is estimated to be between 99.7 and 99.9 per cent alike in the sequence of base pairs that make up the alphabet of the genetic code. But even this scant difference means that any comparison of 1000 base pairs along any particular DNA molecule may turn up between one and three differences. Because most genes are longer than 1000 bases, there may be several genetic differences from person to person in nearly all the 60 000 to 70 000 genes people are now thought to possess.
The vast majority of these differences will be 鈥渘eutral鈥 or will appear in noncoding sequences 鈥 鈥渟ilent鈥 regions of DNA that do not produce any protein product 鈥 and so will not produce any physical effect. Yet among this wealth of variation, new correlations will be found that may appear to bolster the legitimacy of racial categories.
One family of genes already under scrutiny are the so-called HLA genes of the immune system. These genes are responsible for the rejection of unmatched organ transplants and are a rich source of variation: the frequency of one particular variant of an HLA gene has been found to range from 0.2 per cent in a Japanese population to 19 per cent in a group of French whites, for instance. Another is said to be a 鈥渉igh-frequency Chinese marker found in nearly 50 per cent of Cantonese鈥.
鈥淛unk鈥 DNA 鈥 sequences that do not code for any protein 鈥 may be even more informative. One such family of variable sequences 鈥 which differ widely from person to person, for no apparent reason 鈥 are known as VNTRs, or 鈥渧ariable number of tandem repeats鈥. As long ago as 1988, these DNA sequences have been pursued by geneticists as signposts for both disease genes and potential markers of 鈥渞acial鈥 origin.
Of all such DNA 鈥減olymorphisms鈥, the ones most useful for making racial distinctions will be those that are 鈥減opulation specific鈥, or 鈥減rivate鈥, according to Adrian Hill, a molecular geneticist at the Institute of Molecular Medicine in Oxford. While this hasn鈥檛 been true of most DNA polymorphisms described to date, says Hill, 鈥渟ufficient private polymorphisms are turning up to indicate that the genome as a whole must contain large numbers of population-specific variants鈥. Hill also believes that it should be easiest to use DNA polymorphisms to identify African ancestry, since people of African origin seem to have a disproportionate number of 鈥減rivate鈥 polymorphisms.
Forensic tests will not be the only application of such genetic data. Indeed, assessments of human genetic variation may come to have greater political implications outside the criminal justice system. Consider the furore in the US over who qualifies as a 鈥淣ative American鈥. Access to land and social services, as well as considerable social cachet in some circles, now accrues to those who are considered to be descendants of the original inhabitants of what is now the US. But who qualifies as a 鈥渞eal鈥 American Indian?
Blood ties
The US government has long 鈥渃ertified鈥 the identities of Native Americans on the basis of 鈥渂lood quantum鈥. When first introduced in the Allotment Act of 1887, these criteria denied land rights to those who were less than 鈥渉alf-blood鈥, enabling the federal authorities to appropriate millions of acres of 鈥渟urplus鈥 land for the use of 鈥渨hite鈥 settlers. The continued use of blood quantum criteria by the Bureau of Indian Affairs 鈥 now set at 鈥渙ne-quarter blood鈥 鈥 has led to conflicts among Indian peoples competing for benefits tied to federal recognition. It has also forced many tribes to adopt similar racialised 鈥渕embership鈥 policies themselves, overturning traditional ways of creating new kinship ties through intermarriage, adoption or naturalisation.
The US government鈥檚 policy of requiring American Indians to produce official 鈥淐ertificates with Degree of Indian Blood鈥 has also forged alliances across tribal boundaries, and some groups have begun to band together to oppose federal policies and to claim the right of self-determination. In 1985, one woman successfully sued the Bureau of Indian Affairs for denying her federal benefits because she was 鈥渓ess than quarter blood鈥. But this legal precedent has so far failed to move federal authorities to abandon their blood quantum criteria.
Political debate surrounding Indian identity has now intensified, in the wake of the Indian Arts and Crafts Act of 1990. This law makes it a crime to publicly identify oneself as American Indian when selling artwork, or for a gallery to exhibit art as 鈥淚ndian鈥, if the artist cannot provide federal certification.
鈥淪ince the passage of the bill, mediocre but certified 鈥業ndian鈥 artists have started denouncing non-federally certified Indian artists whom they perceive as competitors,鈥 says Annette Jaimes at the University of Colorado. She claims that the implications of the arts and crafts bill go far beyond the Indian art market. 鈥淭he self-ordained Indian identity police have begun to spread rumours and spurious allegations on university campuses and in other institutions about persons whom they claim are impostors 鈥榤asquerading as real Indians鈥.鈥 Those on their hit list are accused of 鈥渆thnic fraud鈥, says Jaimes, with job opportunities, professional credibility and personal integrity under threat.
Political minefield
Into this political minefield may come genetic data on the frequencies of particular genetic sequences in those who claim to be North American Indians. The result could be new criteria for including and excluding individuals as 鈥渞eal鈥 American Indians 鈥 new ways of deciding who 鈥渂elongs鈥 within that categorisation and who does not. The fact that any such genetic test would necessarily be arbitrary 鈥 based on the possession of a limited number of genes chosen on a statistical basis as 鈥渃haracteristic鈥 of Native Americans as a whole 鈥 might not lessen its appeal to federal authorities, and to some members of the North American Indian community as well. Already, geneticists in Ottawa have attempted to distinguish between 鈥淐aucasian鈥 Americans and Native Americans on the basis of a variable DNA region often used in DNA fingerprinting.
Will the same sorry tale be repeated across the globe when the 鈥渉uman genome鈥 is an open book and we can all read our ancestral origins in our genes? Few contemporary commentators doubt that the fruits of the human genome project will, in time, radically restructure our perception of what and who we are. It鈥檚 easy to imagine molecular biology both destabilising existing boundaries between races and defining new ones, perhaps creating new hierarchies based on 鈥渟uperior鈥 and 鈥渋nferior鈥 genes in the process. Easy too to imagine that data on genetic variation might one day become a novel tool of repression in regimes intent on genocide or 鈥渆thnic cleansing鈥.
Yet, used in a different way, data from the human genome project could enhance our understanding of what it is to be 鈥淓uropean鈥 or 鈥淏ritish鈥, 鈥渨hite鈥 or 鈥渂lack鈥. They could even encourage us to begin celebrating our genetic inter-mingling. This is what Salman Rushdie, who describes his novel The Satanic Verses as 鈥渁 love-song to our mongrel selves鈥, urges us to do. After all, says Rushdie: 鈥淢茅lange, hotchpotch, a bit of this and a bit of that is how newness enters the world.鈥