Saxons, Vikings, and Celts

by Bryan Sykes

  For the first six weeks of life, there is no visible difference between male and female embryos. At about that time, the sex gene on the Y-chromosome switches on. This sends a signal to a whole series of other genes situated on other chromosomes, which, between them, actively divert embryonic development away from female and towards male. Embryos that don’t have a Y-chromosome just carry on along the normal female development pathway and are born girls. The X-chromosome has nothing to do with it. Men truly are genetically modified women.

  This mechanism for deciding sex which humans have inherited from their distant mammalian ancestors creates the second of our guides to our genetic origins. Men carry both an X- and a Y-chromosome in all of their cells–except mature sperm. Sperm occur in two different genetic forms, indistinguishable under the microscope and in their swimming capabilities. Stem cells in the male testis are dividing furiously to keep up the supply of sperm and like the other cells in the body have the XY combination of sex chromosomes. At the final division, the cell divides one last time but the resulting sperm only get one of the sex chromosomes, not both. Half the sperm receive an X-chromosome from this division while the other half get a Y-chromosome. The sex of the child entirely depends on which sort of sperm wins the race to the egg. If it’s got an X-chromosome then the egg, which already has one X-chromosome, becomes XX after fertilization, develops as a female embryo and is born a girl. If, on the other hand, the winning sperm contains a Y-chromosome, the fertilized egg becomes XY and develops into a boy. The simple conclusion is this: Y-chromosomes get passed down the male line from father to son.

  Looking backwards, if you are a man, you got your Y-chromosome from your father, who got it from his father. Who got it from his father. Sounds familiar? It is the mirror image of the inheritance pattern for mitochondrial DNA. The Y-chromosome is the perfect complement to mDNA, telling the history of men. But does it have enough genetic variability to be practically useful? It took a very long time to find any mutations at all on the Y-chromosome. For those scientists involved, and thankfully I wasn’t one of them, it was a frustrating few years. In one of the first studies looking for diversity among human Y-chromosomes, 14,000 bases were sequenced from twelve men from widely scattered geographical localities. Only a single mutation was discovered. Another lab sequenced the same 700-base segment from the Y-chromosomes of thirty-eight different men and didn’t find a single mutation in any of them. At long last, and helped by an ingenious technique for finding the elusive mutations, the Y-chromosome began to show its genetic jewels. Slowly, slowly, mutations that had changed one DNA base to another were teased out of the otherwise barren desert of uniformity.

  With these two pieces of DNA we have the perfect companions for our exploration of the genetic past. One follows the female line, the other tracks the male genealogy. What could be better? They had been my guides in Polynesia and in Europe and I knew them well. Among their many qualities is that they both group people into clans. When my colleagues and I had been trying to make sense of the mDNA results from Europe in the early 1990s, we noticed that the 800 or so samples from volunteers from all over Europe fell into seven quite distinct groups based on their mDNA sequences.

  Unlike the chromosomes in the cell nucleus, which are straightforward linear strings of DNA, mitochondrial DNA is formed into a circle, which is a hangover from when the mitochondria themselves were free-living bacteria. The human mitochondrial DNA circle is exactly 16,589 DNA bases in length, but fortunately it is unnecessary to read the entire sequence. Most of the mitochondrial DNA circle is taken up with genes that code for the enzymes involved in aerobic metabolism, which is the prime function of mitochondria in the cell. Because these enzymes have a very particular structure, decided by their amino-acid sequence, mutations in the genes which alter the amino-acid sequence almost always diminish or destroy the enzyme activity. The individuals who are unfortunate enough to experience these mutations in their mDNA usually die. Aerobic metabolism is such a vital part of life that we cannot tolerate even the slightest malfunction. The genetic result is that because these individuals rarely live long enough to have any children, the mutations are not passed on to future generations. If all mDNA mutations behaved like this, we would never find any genetic differences between individuals and it would be quite useless as a guide to the past because everybody’s mDNA would be the same. However, fortunately for our purposes, not all mDNA does code for these vital metabolic enzymes.

  Approximately 1,000 of the 16,589 DNA bases in the mDNA circle have a different function altogether, one that does not depend on the precise sequence. This stretch of DNA is called the ‘control region’ because it controls the way mDNA copies itself during cell division. Fortunately for us, part of this control region comprises a stretch of 400 bases whose precise sequence is unimportant. It is really just a piece of genetic padding. It must be there and it must be 400 bases long for the control region to work properly, but it doesn’t seem to matter what these 400 bases actually are. This is the complete opposite to the parts of mDNA that code for the metabolic enzymes, which, as we have seen, need to have a very particular sequence. The vital consequence for us of this tolerance in the DNA sequence of the control region is that when a mutation happens it doesn’t affect the performance of the mitochondria at all. Instead of killing the individual who carries it, the control-region mutations just carry on unnoticed through the generations, and we can find them.

  During our work in Europe it was the mDNA sequences that we found in the control region that showed us that there were seven principal groups. Within each group, everybody shared a particular set of control-region mutations. The notation that we used to describe these mutations was as simple as we could make it. We chose one particular sequence as our ‘reference sequence’. If we use the metaphor of DNA as a word, then the reference sequence is its standard spelling. The sequence we chose as the standard was the one we most frequently encountered in Europe. If a particular mDNA sequence differed from the reference at the 126th base of the 400 in the control region, then it was denoted simply as 126. If there was another mutation at the 294th position, then the notation became 126, 294. We found a lot of people who shared this particular combination of mutations and they formed one of our seven groups. In other groups there were different sets of ‘signature’ mutations. However, within the groups like the one defined by mutations at 126 and 294, there were plenty of other mutations as well. While about a third of people within the group had just the bare minimum of 126, 294, the rest had one, two, three or even more additional mutations.

  By looking for the signature mutations it was fairly easy to place any individual DNA into one of the seven groups. Occasionally we would find individuals where one of the signature mutations had changed back to the original reference, but on the whole it was quite straightforward. But what did these groups actually signify? It had to mean that everyone within the same group must be related to one another through their matrilineal ancestors, which was the line we were following with mDNA. If two people in the same group had been able to follow their maternal ancestry back in time through their mothers and their mother’s mothers and so on, at some point they would converge. There would have been a woman living in the past who was the common ancestor of both of them. It then struck me, after what now feels like an embarrassingly long time, that if this worked for two people in the same clan it must, by an inevitable logic, also work for the entire clan. If one were to trace back all the maternal lines of everybody within each clan, they would end up with just one woman. There was no alternative. Amazing as it sounds, this has to be true.

  I realized at once that these clan mothers, as I called them, were not some kind of theoretical ancestors, but real living, breathing women. No, not just women, they were mothers as well. Mothers who had survived and whose children, or at least whose daughters, had survived and who in turn had survived and had daughters and so on, right down to the present day. Though men have mDNA, they do not pass i
t on to their children, but they do inherit it from their mothers. Originally to emphasize to myself that these clan mothers were real individuals, I gave them names, each of which began with the letter by which the seven different groups were by then known among scientists. So the clan mother of Group H became Helena, T became Tara, J became Jasmine, X became Xenia, V became Velda, K became Katrine and U became Ursula. Over 95 per cent of native Europeans are in one of the seven maternal clans, and so it followed that these seven women were the maternal ancestors of almost all Europeans. As soon as I had given them names, they came alive and I had to know more about them. I became quite desperate to build up a picture of their lives. I wanted to know all there was to know about these seven women, the women who soon came to be known as the Seven Daughters of Eve.

  The first thing I wanted to know was how long ago these seven women had lived. Were we talking about hundreds, or thousands, or tens of thousands of years ago? The answer came by looking at the extra mutations within the clan. Taking the clan defined by the signature mutations at 126 and 294, which is the clan of Tara and the one to which I belong, everyone within the clan shares these two mutations, for the simple reason that Tara herself had these mutations and everyone in the clan is one of her direct matrilineal descendants. These two mutations have come down through the generations unchanged from the clan mother herself. But how many generations? How long ago did Tara live? That is where the additional mutations come in. Although roughly a third of people in Tara’s clan have only these two mutations, the rest have additional changes. I have one extra mutation, at position 292, which makes my mDNA sequence 126, 292, 294. Other members of the clan have experienced more mutations. All these additional mutations must have occurred since Tara’s time. Fortunately we know the mutation rate for the mDNA control region. It is approximately one change every 20,000 years. Since mutations happen completely randomly, not every line of descent from Tara will experience the same number of mutations. Some may be spared altogether and retain just the signature mutations at 126 and 294. Some maternal lines, like mine, will have been hit once since Tara’s time, others more than once, some not at all. By working out the average number of additional mutations within the clan, we can then estimate how old the clan is, or, to put it another way, how long ago Tara herself lived. For her clan, the average number of additional mutations within the clan is almost exactly 0.85. With a mutation rate of 1 change per 20,000 years, the conclusion is that Tara lived 17,000 years ago.

  Repeating the same calculations for the other six clans, we arrive at estimates for the ages of the other clan mothers. The clan with the greatest number of additional mutations on top of the clan mother’s signature sequence is Ursula’s. Hers is therefore the oldest of the seven clans. The average number of extra mutations in the clan is 2.75, and factoring in the mutation rate, this means that Ursula herself lived 45,000 years ago. Xenia is the next oldest at 25,000 years, Helena next at 20,000 years, then Velda and Tara both at 17,000 years, Katrine slightly younger at 15,000 years and finally Jasmine at 10,000 years ago.

  Working out how long ago these women lived was a big step to discovering what their lives were like. Now I knew when they lived, could I discover where? I used three tests to find out. First, knowing the current whereabouts of the clan throughout Europe, I discovered where the clan was concentrated, reasoning that even after so many thousands of years, this might still be close to its origin. However, more important was to plot where the clan had accumulated the most additional mutations. The reasoning here was that the clan would have had longest to ‘age’ close to its origin, where the clan mother herself lived. To give you an example, the clan of Velda reaches its highest frequency in two places–northern Spain and among the Saami of northern Scandinavia. But it is far more varied, in the sense that it has accumulated far more extra mutations, in Spain than in Lapland. So I placed Velda herself in northern Spain, rather than in the far north of Norway and Sweden. Which brings me on to the third test. The location of the clan mother has to have been habitable at the time. In Velda’s case, we know from the archaeological records that people were living in northern Spain 17,000 years ago, the date estimated from the additional mutations in the clan, but they were certainly not living in northern Scandinavia, which was under several kilometres of ice. By the same process, the other clan mothers were located to Greece (Ursula), the Caucasus mountains (Xenia), southern France (Helena), northern Italy (Katrine and Tara) and finally Syria in the Middle East (Jasmine).

  With information from climate records and the archaeological evidence, I was able to find out what conditions must have been like for these women living at these locations at those times in the past. I discovered what their landscape was like, what sort of diet they had, what age they reached and, armed with this information, I wrote imagined lives for them.

  Since they were published, the response has been both surprising and intriguing. My laboratory was overwhelmed by requests from all over the world from people who wanted to know from which of these women they were themselves descended. We had already repeated the process worldwide and found a total of thirty-six equivalent clans, so we could deal with requests from anywhere. We could not possibly handle this demand in the lab, if only because we were prevented from carrying on any commercial activities by the rules of our principal sponsors, the Wellcome Trust. So the University rapidly formed a spin-off company, Oxford Ancestors, to perform this service. But that is of only passing interest compared to the quite extraordinary underlying emotion that the concept clearly aroused. It proved to me that to many people, of which I am one, the idea that within each of our body cells we carry a tangible fragment from an ancestor from thousands of years ago is both astonishing and profound. That these pieces of DNA have travelled over thousands of miles and thousands of years to get to us, virtually unchanged, from our remote ancestors still fills me with awe, and I am not alone. One unexpected effect is that when two people discover that they are both in the same clan, they really do feel like close relatives, like cousins or siblings. I have seen this happen time and again, and indeed on the Oxford Ancestors website one of the most popular activities is discovering genetic relatives and then swapping personal information and often finding uncanny similarities of personality and circumstance. Even if this is all retrospective wisdom, after the test rather than before, the strength of feeling is very strong. There are even Jasmine parties organized by members of the clan.

  I recently tested the DNA of our Vice-Chancellor, the executive head of Oxford University–I rarely travel anywhere without a DNA sampling brush–and discovered that he and I are not only in the same clan of Tara, but have exactly the same mDNA sequence 126, 292, 294. This means that as well as a common ancestor 17,000 years ago in Tara herself, we must share a much more recent maternal ancestor. I don’t know who that is, but the point of the story is that, for better or worse, I feel now very differently about the Vice-Chancellor. So much so that, were we to have a severe disagreement, it would be hard for me to take it quite so seriously. It would be like arguing with my cousin.

  A few years later, the same treatment became possible for the Y-chromosome. The details of the genetic changes were slightly different, and we will see how in a later chapter, but the principle remains the same. Whereas there are seven maternal clans which predominate in western Europe, there are only five principal paternal clans defined by the Y-chromosome. Each of these began with just one man, but for reasons that will become clear, it is much harder to know when and where they might have lived.

  7

  THE NATURE OF THE EVIDENCE

  The collection phase of the Isles research project began ten years ago, in 1996, under the title of the Oxford Genetic Atlas Project. I obtained ethical permission to collect DNA samples from volunteers with the specific objective of discovering more about our genetic history. Over the next few years, my research team and I worked our way all over the Isles. We collected over 10,000 DNA samples and travelled over 80,000
miles by train, plane, boat, car and bus. Eventually I had to draw a line under the collection phase and concentrate on distilling some meaning from the thousands of DNA samples that now lay crowded in the lab freezers. We had been putting them through the analytical procedures more or less as they were being collected, converting the drab white threads of DNA into the sequences which would, or so we dearly hoped, hold the secrets of the ancient people of the Isles. Displayed on a computer screen they looked detached, dead–nothing like the talismans of ancient histories that I hoped they would become.

  It took a lot of mental effort constantly to remind myself that every single one of these strings of letters and numbers represented the journey of an ancestor. A journey that at one stage almost certainly involved a sea crossing in a fragile craft to landfall on the Isles and an uncertain future. Fantastic though it sounds, it had to be true that each one of the thousands upon thousands of read-outs that flashed from the analyser to the computer in a fraction of a second had been carried across the sea in the cells of an ancestor. How could I get these mute listings to tell me their stories? How could I get them to sing? If only, I thought one day, I could read in the letters of the genetic code the language of the bearer. How wonderful that would be–and how much easier than the task that lay ahead. If, just by looking, I could recognize a Gaelic word or a Saxon spelling somewhere in the sequence of DNA letters. But the genes were stubbornly silent, oblivious to the tongues of their bearers.

  Mathematicians have devised a whole array of statistical tests to sieve through DNA results, mechanically and without feeling. Indeed, most scientific papers on this kind of genetics spend at least half the time agonizing over what is the correct statistical treatment. It is necessary, if only to get results published, to know how to do this and fortunately we had in the lab several people skilled in the art. They, in particular Eileen, Jayne and Sara, put the accumulating genetic data through their paces. They ran Hudson tests, Mantel tests, distance-based clustering analyses, drew genetic matrices based on Fst and Nei’s D, performed spatial auto-correlation tests and many more. Here are some of the results that came screaming out of the computer. It is a set of genetic comparisons from mitochondrial DNA between the four regions of the Isles.