by Bryan Sykes
On my imaginary map, I moved all the Irish gene-coins across the Irish Sea and began to distribute them according to their geographic origins. Each one of these was the end point of a journey – the journey of a line of ancestors stretching through maternal and paternal threads way back into the deep past. We know, from the archaeological records, that every one of these ancestral journeys must have ended in Ireland within the last 9,000 years when the first Irish built their timber-framed houses on the banks of the River Bann at Mount Sandel. Before that, as we know, Ireland had been uninhabited since the Ice Age. Could it be that some of these DNA fragments from today’s Irish men and women have actually been there all that time? How would we know?
I began with the maternal DNA. I knew, from the identity of their clan mothers, when and where all these journeys had begun. And from the locations on the map of Ireland, I also knew where these journeys had ended. Well, not ended, because many of these genes will go on travelling the world for millennia to come. To be more accurate, I knew where these immortal time travellers had reached by the late twentieth century.
The longest journeys, in both time and distance, had been travelled by Ursula’s descendants. Ursula herself had lived in Greece about 45,000 years ago, at the start of the Upper Palaeolithic, and had shared the land with the far more ancient Neanderthals. We knew this date from counting up all the mutational changes that had happened among her descendants and dividing this figure by the mutation rate. As we have seen, among the seven principal clans in Europe there were far more changes in the descendants of Ursula than in any of the others. That simple fact meant the clan of Ursula was the oldest of the seven. We arrived at the age of the clan, and thus the time in the past when Ursula herself lived, by factoring in a mutation rate of one change in every 20,000 years. From thousands of DNA samples from all over Europe we knew that, on average, Ursulans have 2.25 mutations in their mitochondrial DNA compared to the DNA of Ursula herself. That puts the age of the clan at 2.25 × 20,000 = 45,000 years.
We arrived at Ursula’s own location in Greece by looking to see where the clan today is both the most frequent and at its most diverse. So long as there was good archaeological or climatic evidence that the location was inhabited, or at least habitable, at the time, then that was where we placed the matriarch. I realize from many letters that it is a frequent and understandable misunderstanding (if there can be such a thing) that we have located the skeleton of Ursula and the other matriarchs and then worked out how long ago they lived from carbon-dating. But this is not so; it is all accomplished by reconstructions.
From the DNA fragments now displayed on the map of Ireland, I could see that almost exactly 10 per cent of Irish men and women are the direct maternal descendants of Ursula, each carrying her DNA modified by the occasional mutation. Converting that proportion of 10 per cent to actual numbers of people means that of a total population of 5.7 million there are roughly 570,000 Ursulans in all of Ireland. Altogether, counting Dan’s published data, we have the mDNA sequences of 91 Ursulans from a total of 921 Irish samples. However, because the clan is so old and there has been such a long time for mutations to accumulate, we found only three people who have Ursula’s sequence unmodified by genetic change. As all three are customers of Oxford Ancestors, I can trace their present whereabouts and none of them lives in Ireland! One lives in Hampshire in southern England, another in London and the third in New York. It is their grandmothers and great-grandmothers who lived in Ireland and who then joined the stream of emigrants in the nineteenth and early twentieth century. But their DNA did pause in Ireland before it resumed its journey around the globe.
I shall refer to these three as ‘pure’ Ursulans, while, I hope, avoiding the implication that mutations in the others have somehow sullied the pristine genetic heritage of Ursula herself. I hope I am not accused of implying that their DNA is, in contrast, somehow impure, which is complete nonsense of course. But it has been modified. Of the other Irish Ursulans, some have one change on the ‘pure’ Ursulan background – and I call them first-generation Ursulans. There are 22 of them. By the same token, there are 23 second-generation, 26 third-generation, 10 fourth-generation, 5 fifth-generation, one sixth-generation and one seventh-generation Ursulan. The third-generation Ursulans, with three mutations compared to Ursula herself, are the most frequent, closely followed by Ursulans of the second ‘mutational’ generation with two changes compared to the original. The numbers in higher generations tail off slowly until we reach the Irish Ursulan record-holder, a doctor now living in Chichester, in Sussex, with seven changes since Ursula. I had better stress once again that the ‘pure’ Ursulans and all the others up to and including the record-holder are separated from Ursula by the same 45,000 years and, roughly, the same number of actual maternal generations. It is chance alone that has left the three ‘pure’ lines untouched by mutations while the good doctor’s has been hit seven times.
We are in the most technical part of the book and I beg for your indulgence to explain a very important point. From the numbers of ‘pure’ and first-, second- and higher-generation Ursulans, we can work out the average number of mutations over all the Irish Ursulans. It comes to fractionally over 2.5. If we now multiply this by the mutation rate of one change in every 20,000 years, it comes to just over 50,000 years, which is older than Ursula!
Actually, this is not that far from the 45,000-year date for Ursula herself and well within the mathematical error of the estimate. But it is an awful lot longer than the 9,000 years we know that people have been in Ireland. How can we explain this apparent discrepancy? We have 50,000 years’ worth of accumulated mutation in an island which we know has only been inhabited for 9,000 years. It has to mean that most of the mutations in the Irish Ursulans must have already occurred before and not after they arrived in Ireland. We cannot just use the 50,000-year genetic date and say that is when Ireland was first inhabited.
My colleague Martin Richards and I got into a lot of trouble when we first used a superficially similar argument to back up our controversial proposition that the ancestors of most Europeans were Palaeolithic hunter-gatherers who had arrived a long time before the Neolithic farmers. We said that there was far too much accumulated mutation in all the major European clans, save Jasmine, to have developed in the 10,000 years or so since farming had been invented in the Middle East, and therefore these clans were Palaeolithic in origin. The counter-argument was vigorously expressed in terms of a Martian metaphor. Suppose a representative selection of Europeans had been transported to Mars and then, a few years later, had their DNA sampled and analysed – presumably by Venusians who didn’t know about the landing. They had then done the calculations and showed that, according to the amount of accumulated genetic differences, the Martians had been there for tens of thousands of years, whereas we know they had arrived only a few years earlier. The flaw in the Venusians’ argument – and by implication in ours – was that they had assumed that all the mutations had accumulated after the earthlings arrived on Mars, when, in fact, they had all occurred before they set off from Earth.
Although Martin and I, for a number of reasons, did not think this was a good analogy for what we had actually done, we none the less set out to try to prove that the mutations in Europeans (we are back on Earth now) had accumulated in Europe and had not been imported from the Middle East by Neolithic farmers. Martin particularly, helped by our new recruit, theoretical physicist and mathematician Vincent Macaulay, spent three long years doing this.
To cut a very long story short, they scoured the Middle East for as many DNA samples as they could find, then searched these for matches to the DNA from Europe. The point was to find out how many of the mutations in Europe were genuinely European and how many had already happened in the Middle East. Basically, if a match was found, and if we could be certain that it had actually originated in the Middle East, rather than being carried back there from Europe by some sort of reverse migration, we subtracted it from our tally of �
��European’ mutations and did the time calculations once again. It was an exhaustive, and exhausting, analysis which in the end gave us a set of dates for the settlement of Europe that we could all rely on. Fortunately, they were not so very different from our original ones and did not reverse our conclusion that most Europeans had hunter-gatherer ancestors.
Turning back once more to the Irish Ursulans, could I do the same sort of thing to work out which mutations were Irish ‘originals’ and which were imported? If I could, then the genetic dates would mean something. This I did by checking each of the Irish Ursulans’ DNA sequences against every other sequence that I knew about from all over, first Europe, and then the world. I was looking to see how many of them had also been seen outside Ireland. Although my computer helped a great deal in sorting all the results so that identical sequences appeared on consecutive lines on the screen, I also checked the list of Irish Ursulans one by one. One Irish Ursulan, a lady from Donegal, for example, had only one matching sequence and that was a man from the Czech Republic. It is very unlikely we will ever know the precise tracks that trace the wanderings of the ancestors of these two genetic relatives back to the woman whose DNA they both share. But it is in just these tracks, like footprints in the sands of time, that we can read the signals from the past.
I found, at the end, that out of the 91 Irish Ursulans, 68 had matching sequences elsewhere. Only 23 were unique to Ireland. In many cases I could find their immediate predecessors, in a genetic sense, within Ireland. So a fourth-generation Ursulan sequence, for example, would usually have a third-generation sequence nearby. In these cases I thought it was reasonable to regard that fourth mutation as having happened in Ireland. Counting up these home-grown mutations and factoring in the mutation rate as usual gave me a corrected date for the clan in Ireland of a little over 7,000 years – 7,300 to be precise. This was much more reasonable than the 50,000 years which counted all the Ursulan mutations as if they had all happened in Ireland.
Genetic dates, like the 7,300 years for the Irish Ursulans, are not very accurate. They are estimates. We find this concept particularly difficult to grasp because we are accustomed to dates being very precise. The 7,300-year date for the arrival of the Irish Ursulans is an estimate. The date could vary a thousand years either way and still fall within the scope of the estimate. Forgetting the inaccuracies for the moment, what does this date mean? It is an estimate for the length of time it would have taken for all the Ursulan mutations to have accumulated within Ireland. If the ancestors of all 91 Irish Ursulans had arrived at the same time and their mDNA mutations had accumulated since then, the genetic estimate for their arrival would have been 7,300 years ago. Of course, it is very unlikely indeed that they all arrived at once. Some would have come more recently, but in that case, to achieve the average figure, others must have arrived more than 7,300 years ago to balance out the more recent arrivals.
I went through the same procedure with all the other Irish maternal clans, checking to see in each one how many looked as though they had mutated to their final form in Ireland and not elsewhere. From there I calculated the clan arrival times in the same way as I had for the Ursulans. All of them came out between 7,500 and 4,500 years ago. Ursula was still the oldest clan in Ireland and, in common with the rest of Europe, Jasmine was the youngest. It was Jasmine’s clan that Martin Richards and I had linked to the arrival of Neolithic farmers in Europe from the Middle East. The others clustered around the 5,000–6,000-year average. Even bearing in mind the approximate nature of these genetic dates for the settlement of Ireland by the various maternal clans, they are all way before the time, around 200 BC, when the Iron Age Celts were supposed to have arrived. It was beginning to look as if the ancestors of today’s Irish had been there for a lot longer than anybody thought.
But within Ireland, when I looked at the maternal clans in the different provinces of Ulster, Leinster, Munster and Connacht, there was very little noticeable difference between any of them, though the numbers in each were too low to be sure of statistical significance.
If women had been in Ireland for a very long time, what about the men? Just as mitochondrial DNA traces our maternal ancestry, so the Y-chromosome follows paternal genealogies. Like mitochondrial DNA, Y-chromosomes also experience random mutations over the course of time. The precise nature of the mutations might be different between mDNA and Y-chromosomes, as we shall see, but the principles are the same. The accumulation of mutations along paternal genealogies over a very long time means that there are now tens of thousands of slightly different Y-chromosomes which can be distinguished by genetic tests. If two men have the same Y-chromosome fingerprint, then they have usually inherited it from a common patrilineal ancestor. That’s exactly the same principle as saying that if two people have the same mitochondrial DNA sequence they have inherited it from a common maternal ancestor. Although Y-chromosomes are quite different from mitochondrial DNA in the way they change genetically, that doesn’t matter so much when it comes to interpreting the signals they are bringing us from the past.
Just as each of us belongs to one of a small number of maternal clans, so men can be assigned to a paternal clan by the genetic characteristics of their Y-chromosome. From research done throughout the world over the past decade, Y-chromosomes can now be separated into twenty-one paternal clans, eight of which occur in Europe. Of these eight clans only five occur in the Isles to any appreciable extent. Following the tradition of the maternal clans, I have given them names. They are the clans of Oisin, Wodan, Sigurd, Eshu and Re. Like the maternal clans, each founded by a matriarch, the paternal clans must, by the same logical inevitability, also have been started by a single man – the clan father or patriarch. Every man within a clan is a direct paternal descendant of the clan father and has inherited the patriarch’s Y-chromosome, modified by mutations over the intervening millennia.
The different paternal clans are told apart by single DNA changes, just like the mutations in mitochondrial DNA. However, these Y-chromosome sequence changes occur far more slowly than they do in mDNA. There is usually only a single DNA sequence mutation between the Y-chromosomes of one clan and another, even though they have been separated for tens of thousands of years. Luckily for us, the Y-chromosome also experiences a second, much swifter, type of mutation that can split a paternal clan into hundreds, if not thousands of separate paternal lineages. These fast mutations happen, like all DNA changes, when cells divide and there is an error in the usually immaculate copying mechanism.
Along the Y-chromosome are patches of DNA sequence that, when looked at closely, consist of reiterated blocks of short sequences. Treating DNA sequences like a word, these are the genetic equivalent of a bad stammer. It is as if an otherwise smooth read-through just gets stuck. Take the four DNA letters TAGA, an outwardly unremarkable sequence. For some reason, TAGA tends to trip up the DNA copying mechanisms on parts of the Y-chromosome where it is repeated a number of times. Cells can handle a few repeats. The double reiteration TAGATAGA causes no difficulties. Even ten repeats, one after the other, is manageable. But after that the stammering begins. After twelve repeats cells find the blocks of TAGA very difficult to copy accurately and make mistakes much more often than they normally would with regular sequences. What they do wrong is to add an extra TAGA, or forget to copy one. So a Y-chromosome with, let’s say, fourteen TAGA repeats mutates into one with fifteen repeats. Because the cell has such trouble with copying this type of stammering sequence accurately, the rate at which these mutations occur is hundreds of times faster than the regular type of spelling-change mutation, where, for example, a C changes to a T. It is an even faster rate than mDNA, with its comparatively lax error-checking mechanisms.
There are dozens of places along the Y-chromosome where these tricky stammering segments are to be found and they can be used in combination to define tens of thousands of different Y-chromosomes. Because Y-chromosomes, alone among the nuclear chromosomes, are not shuffled at each generation, the combina
tions can persist for a very long time, changing only when another mutation occurs at one of them. In this respect, the Y-chromosomes can be interpreted just like the mDNA with first-, second- and third-generation mutations changing the Y-chromosome fingerprint of the ‘pure’ patriarch. And, just like mDNA, the number of mutations can be added up to get an idea, again only approximate, of time passing. In reality, because these mutations happen so quickly, in relative terms, it is hard to know what the clan patriarch’s Y-chromosome fingerprint actually was. But as we shall see, this is not really a problem when we look in detail at the Y-chromosome of Ireland.
In my mind’s eye, I collect up the maternal gene-coins from the imaginary map of Ireland, move them to one side and begin to distribute the paternal Y-chromosome equivalents in their place. As soon as I look at the Irish pile, one thing stands out. The vast majority of Irish Y-chromosomes are members of just one clan, the clan of Oisin. It is precisely because of its predominance in Ireland that I gave the clan this name. Oisin was the son of the hero of another of the Irish mythical cycles, Finn mac Cumhaill, sometimes transcribed as Finn mac Cool. Finn is the leader of the Fianna, a band of warriors chosen only after an appropriately gruelling selection process. His son Oisin, meaning Little Deer, is bewitched by Niav of the Golden Hair, the daughter of the Underworld king who reigns over Tir na n’Og, the World of the Forever Young. Oisin goes to this other world to be with Niav and spends his life writing poetry and songs. Eventually he becomes homesick and is eager to visit his own world once more. Niav warns against this, but Oisin is adamant and sets out, though he promises to heed her warning not to set foot on Irish soil. He plans to avoid disobeying Niav’s instructions by riding everywhere on a horse and so not touching the soil. In an extreme version of Back to the Future, Oisin realizes when he returns to Ireland that 300 years have passed while he was relaxing in Tir na n’Og. The shock of this discovery makes him fall from his horse and, as soon as he touches the ground, he instantly ages 300 years and crumbles into dust.
