Are DNA tests the missing link?

Oprah Winfrey says she has Zulu blood. She's not alone. Thousands of people now pay for DNA tests to uncover their roots. But is it really possible to pinpoint your ancestry?

When scientists announced the completion of the Human Genome Project at a press conference in June 2000, they went out of their way to extol the unity of humankind. We all share more than 99.9 per cent of our DNA, they said. Yet five years later the human genome is being used to identify our racial differences and geographical origins - sometimes with dubious precision.

When scientists announced the completion of the Human Genome Project at a press conference in June 2000, they went out of their way to extol the unity of humankind. We all share more than 99.9 per cent of our DNA, they said. Yet five years later the human genome is being used to identify our racial differences and geographical origins - sometimes with dubious precision.

Several companies have sprung up offering testing services that claim to be able to trace our genetic roots. In Britain they offer to identify which of seven ancient European clans we are descended from. In America they promise to find out whether a person has Jewish, African, European or native-American genes. Oprah Winfrey, the chat-show queen, has even had a test. "I went in search of my roots," she says, "and had my DNA tested, and I am a Zulu."

Sometimes the aim of those wanting a test is to fill a void in their personal history - which can be especially painful for the descendants of the African slave trade. Others have wanted to complete a genealogical search of their family, a hunt that may have come to a dead-end using conventional paper trails. Some Americans have even paid for DNA tests in the hope of financial gain. If they can prove they are descended from certain Native American tribes they can claim a share of profits from casinos on tribal lands.

Britain's watchdog on genetics, the Human Genetics Commission, says some claims made in the name of "genetic genealogy" can be misleading. "The scientific information that genealogy tests could provide might not be as precise as some of the companies suggest, and this was an area in which people should be aware of the reality of what they are being offered," the commission says.

Tracing family roots has traditionally relied on using genealogical records including certificates of births, marriages and deaths, which allow the tracing of relationships using surnames that are passed down through the male line. But the recent advances in DNA analysis have offered a new possibility of tracing our ancestors through the sequence of chemical "letters" that make up the alphabet of our genetic code. If two living people share a similar DNA sequence, it could mean that they once shared a common ancestor.

Human DNA is stored in 46 chromosomes arranged in 23 pairs, yet just one of these chromosomes is particularly useful in tracing origins. All chromosomes except the male Y-chromosome engage in swapping, or recombination, of DNA fragments between each chromosome in the pair. This mixing produces a patchwork of ancestral DNA on a single chromosome, making it hard to decipher your ancestors based on analysing this DNA alone.

A much better tool, for men at least, is the Y-chromosome that determines maleness. Men inherit just one such chromosome from their fathers and, being on its own in the cell, it does not get involved with the messy business of recombination. A boy's Y-chromosome is more or less identical to that of his father, his paternal grandfather, great-grandfather and so on. The chromosome does change over time, but the process is very slow.

Analysing the genetic sequence of the Y-chromosome is therefore an excellent tool for looking at male ancestry. In fact there is now a thriving business doing just this, especially for Americans trying to trace their family roots in Europe. Two American companies, DNAPrint Genomics, based in Sarasota, Florida, and Family Tree DNA of Houston, Texas, both offer a large selection of DNA tests that attempt to locate a person's racial or ethnic origin or to find genetic factors shared by people with the same surnames. Both use tests based on analysing the Y-chromosomes of men.

The problem, says Professor Mark Jobling of Leicester University, who studies the genetics of the Y-chromosome, is that although such tests may indicate a degree of relatedness, they do not provide accurate information on how far back the common ancestor lived. "Dating is limited. Saying that two men share recent patrilinear ancestors is possible, but showing when they shared that ancestor is more problematic."

Another useful tool for tracing human origins is mitochondrial DNA. This is the only DNA that occurs outside the cell nucleus, and is passed on from mothers to both their daughters and sons. Again there is no recombination to mix up the mitochondrial DNA sequence, allowing scientists to trace ancestors back many generations. But unlike the Y-chromosome, both men and women have mitochondrial DNA, which permits ancestral tracing for both sexes.

Just as the Y-chromosome has shed light on the early evolution of men, mitochondrial DNA has revealed the origins of maternal ancestors going back many thousands of years. In fact scientists have even located a notional "mitochondrial Eve", the last common female ancestor of all humans, who lived some 200,000 years ago.

The same sort of research has shown that there are about 10 broad divisions of mitochondrial DNA in Europeans today. Each category is descended from a single common female ancestor who lived between about 15,000 and 30,000 years ago. Oxford Ancestors, a company set up by Professor Bryan Sykes of Oxford University, offers tests which place people in one of seven mitochondrial DNA "clans" to represent what Professor Sykes has romantically dubbed "The Seven Daughters of Eve". Professor Sykes has even given each founder female a fictional name, such as Ursula and Jasmine, and claims that these women gave rise to "clans" that once lived in certain areas of Europe, such as north-east Italy, the Dordogne valley or the Baltic region.

"It's all part of increasing your sense of belonging in the world at a time when people feel increasingly disconnected," Professor Sykes says. "It's not something you'd think was intrinsically useful." Nevertheless, nearly 20,000 people have each paid about £180 for a DNA test carried out by Oxford Ancestors.

However, not every geneticist is happy with the idea of using mitochondrial DNA to suggest definite ancestral origins for individuals. "This information can tell us about the process of colonisation in order to paint a more detailed portrait of Europe's past but mitochondrial DNA is only a tiny fraction of our genome," says Martin Richards, a geneticist at Leeds University. "Overall, members of a so-called 'clan' are no more likely to be more closely related to each other than to other members of their population. Some people do feel it gives them a feeling of being related, but this is imaginary: there is no mystical bond between them."

Every DNA test used to trace our ancestral origins has limitations. The first is simply based on the fact that if we go back far enough, we are all related to one another. This is because of a simple mathematical fact - the number of our ancestors increases exponentially as we go back in time. Two parents become four grandparents, who become eight great-grandparents and so on. By the time we go back a further 15 generations, each of us has more than half a million direct ancestors. It is hardly surprising that we can find relationships we didn't expect when we analyse our DNA sequence.

The other limitation of a DNA test is that it focuses on a relatively short element of your entire genome. The Oxford Ancestors test may tell you that you belong to the "clan of Ursula", but this only refers to your mitochondrial DNA chain - which is a tiny fraction of a much bigger genome that is descended from thousands, perhaps millions, of other ancestors.

In addition, when companies such as DNAPrint Genomics offer to match your DNA with a racial or ethnic group, the accuracy with which they can do this relies on the quality of their genetic database, which can be very poor. This error was committed by the BBC television programme Motherland, transmitted in 2003, which carried out DNA tests on black Britons whose grandparents came from the Caribbean. The programme filmed these people being reunited with their long-lost relatives in Africa.

One woman from Bristol was shown meeting her "relations" from the tiny island of Bioko off Cameroon. The woman's mitochondrial DNA matched only eight people in the global database and they all came from Bioko, the programme claimed. Unfortunately, scientists subsequently found the same DNA sequence in people living in Mozambique. The Bristol woman was just as likely to be related to these people as the islanders.

"There's a lot of unreliability in these tests because of the quality of the database they rely on," says Professor Jobling. "And it's difficult for companies offering such tests to say that this is going to cost you £150 but it's not going to be worth very much."

This may well turn out to be true for Oprah Winfrey. She believes she's a Zulu from Southern Africa because her DNA test says so. Yet there are few records of Zulus being involved in the Atlantic slave trade, which mainly focused on tribes from west Africa. Whatever the DNA sequence that linked Oprah with Zulus, it may yet turn out to be present in west Africans.

Although DNA tests have proved invaluable in identifying very close relatives - highlighted by paternity tests - they have proved problematic in finding our distant cousins. But what they have shown is that if we go back far enough, we are all related.