by Pbo, Svante
The Dubrovnik meeting ended with a culinary feast that lasted for hours and left us all full of excellent food and pleasantly inebriated. During my career I had been part of many collaborations but none had been as good as this one. Still, I felt a sense of great urgency to bring the project to completion. During dinner, I impressed on everyone that we were now on a tight time schedule, both because the world was awaiting our results after the announcement at the AAAS meeting and because we didn’t know what Eddy Rubin was doing in Berkeley with the Neanderthal bones we knew he had collected. Although I hardly ever have bad dreams, I claimed in my improvised speech at the dinner that I had had nightmares about a paper from Berkeley appearing a week before ours with all the same insights we had found.
The next morning I slept on the plane back to Germany. Shortly after returning to Leipzig, I came down with a cold, which developed into a fever and then chest pains in sync with my breathing. I went to the hospital and was diagnosed with pneumonia and given a prescription for antibiotics. But shortly after I got home I received a call to return to the hospital immediately. The lab results suggested I had blood clots somewhere in my system. I soon found myself staring at a CT scan showing blood clots clogging large parts of my lungs. It was a rattling experience. If these clots had reached my lungs as one large clump instead of several smaller pieces, I would have died instantaneously. The doctors blamed the blood clots on too much flying and perhaps the long, cramped bus trip through the night from Split to Dubrovnik. I was prescribed anticoagulants for six months and began researching therapeutic alternatives with the intensity that only comes from being personally affected. To my amazement I stumbled upon references to my father’s work from 1943. He had elucidated the chemical structure of heparin, the drug the doctors had given me when I entered the hospital and which had perhaps saved my life. While I found this amusing, I was also quite shaken. It threw a stark light on my family background. I had grown up as the secret extramarital son of Sune Bergström, a well-known biochemist who had shared the Nobel Prize in 1982 for the discovery of prostaglandins, a group of natural compounds that have many important functions in our bodies. I had seen him only occasionally during my adulthood, and the fact that he had worked on the structure of heparin was just one of the innumerable things I did not know about him. The sadness I felt for not having known my father made me realize even more strongly that I wanted to be there when my own three-year-old son grew up. I wanted him to know me. And I wanted to see the Neanderthal project through to completion. It was too early for me to die.
Chapter 18
Gene Flow!
________________
We began sequencing our five modern genomes in May 2009. The pristine DNA, free of the bacterial contamination and chemical damage that marred our Neanderthal samples, yielded about five times as many DNA sequences from each of the five people as we had generated from the Neanderthal. Only a year or two earlier, sequencing those genomes in Leipzig had been unimaginable, but the sequencing technologies such as those marketed by 454 and Illumina had now made it possible for small research groups like ours to sequence several complete human genomes in just a few weeks.
Using the approach he had described in Dubrovnik, Ed estimated how long ago the five present-day human genomes had shared common ancestors with the human reference genome. He found that the European, Papuan, and Chinese individuals shared common ancestors with the reference genome a little over 500,000 years ago. Adding the San from South Africa to the group pushed the point of divergence back to almost 700,000 years ago. The divergence between the San (and related groups) and other people in Africa and elsewhere was among the deepest seen between present-day people. This put the 830,000 year age estimate for the common ancestor of the Neanderthal and present day-human genomes in perspective: having diverged only 130,000 years earlier, Neanderthals were different from us, but not by a lot.
Such calculations must be treated carefully, as they give one single value for the age to a common ancestor as if that was true for the entire genome. Genomes are not inherited as units, however, which means that each part of an individual’s genome has its own history and therefore its own common ancestor with the genome of any other individual. This is because each person carries two copies of each chromosome and one of these is independently passed on to a child. So each chromosome has its own independent pattern of history—or its own genealogy, if you will. In addition, each chromosome pair exchanges pieces with each other in an intricate molecular dance called recombination that takes place when egg cells and sperm are formed. Therefore, not only does each chromosome in a population have its own genealogy but each piece of each chromosome does, too. Thus, the ages that Ed had calculated for common ancestors with the reference human genome, 830,000 years for the Neanderthal and 700,000 years for the San, represent grand averages across all parts of the genome.
In fact, when we compared DNA regions from two present-day people with each other, we could easily find regions where they shared a common ancestor just a few tens of thousands of years ago but also regions where they last shared an ancestor 1.5 million years ago. The same was true in a comparison of present-day people with the Neanderthals. So if someone could take a walk down one of my chromosomes and compare it to both a Neanderthal and a reader of this book, that chromosomal pedestrian would find that sometimes I would be more similar to the Neanderthal than to the reader, sometimes the reader would be more similar to the Neanderthal, and sometimes the reader and I would be more similar to each other than to the Neanderthal. Ed’s average simply meant that there are slightly more regions of the genome where the reader and I are more similar to each other than either of us are to the Neanderthal.
It is also important to realize that 830,000 years ago is the average age when DNA sequences in people living today have common origins with the DNA sequences carried by the Neanderthal fossils. At that time, these DNA sequences existed in a population whose descendants would eventually give rise to the ancestors of Neanderthals as well as the ancestors of present-day humans. But this was not the time when the populations that were to become modern humans and Neanderthals split from each other. That must have happened later. The reason for this is that when we trace the history of DNA sequences in a present-day human and a Neanderthal back in time, the two lineages enter the last population ancestral to both modern humans and Neanderthals—the population where the split between the two groups first happened—and then enter the variation that existed in that ancestral population. So the 830,000 years is a composite age that includes both the time when modern humans and Neanderthals were separate populations and the genetic variation that existed in their common ancestral population.
The ancestral population is still totally mysterious to us, although we think it lived in Africa and that some of its descendants eventually left Africa to become the ancestors of Neanderthals. Those who stayed behind were to become the ancestors of people who live today. Estimating when those two groups split using differences in DNA sequences is a tricky proposition, much more so than estimating the time when DNA sequences shared common ancestors. For example, if the population ancestral to Neanderthals and people today contained a lot of variation, more of the DNA sequence differences we found would have accumulated in the ancestral population rather than after Neanderthals and modern humans went their separate ways. This would make the population split relatively recent. We were able to crudely estimate the level of variation in the ancestral population from how different the time estimates to common DNA ancestors were for different segments of the genome. To estimate the population split time, we also needed to know the generation time, or the average age at which individuals produced offspring, something that we obviously didn’t know. Taking these uncertainties into account as best we could, we came to the conclusion that the population split seemed to have happened sometime between 270,000 and 440,000 years ago, although even that might also underestimate the uncertainty. Nevertheless, the ancestors of people tod
ay and the ancestors of Neanderthals probably went their separate ways at least 300,000 years ago.
Having gauged how different Neanderthals and modern humans were, we returned to the question of what happened when the ancestors of present-day people left Africa and met their long-lost Neanderthal “cousins” in Europe. To see whether those modern humans and Neanderthals exchanged genes, Ed quickly mapped our five human genomes to the chimpanzee genome and David and Nick repeated their analyses. I was convinced that the results would now be reliable, and I secretly suspected that the extra similarity between the Neanderthals and the Europeans and Chinese would disappear.
On July 28, I received two long e-mails from David and Nick. It is a testimony to David’s passion for science that the analyses went forward even though his wife Eugenie gave birth to their first child on July 14. Nick had done the ten possible pairwise comparisons among the five modern human genomes. In each case, he identified SNPs where a chromosome in one individual differed from a chromosome in the other. He found about 200,000 such differences between any one pair, more than enough SNPs to accurately determine whether the Neanderthal was closer to one human or the other.
Nick found that the Neanderthals matched the San in 49.9 percent and the Yoruba in 50.1 percent of cases. This was expected since Neanderthals had never been in Africa and therefore should not have more of a relationship to some Africans than to others. When he used SNPs where the French differed from the San, the Neanderthals matched the French in 52.4 percent of cases. We now had such a vast amount of data that there was only a 0.4 percent uncertainty in these values. So it was very clear that the French genome was more similar to the Neanderthals than was the San. For the comparison using the French and the Yoruba, the corresponding value was 52.5 percent. For SNPs where the Chinese differed from the San and the Yoruba, the values were 52.6 percent and 52.7 percent, respectively, and for SNPs where the Papuans differed from the Africans, it was 51.9 percent and 52.1 percent. When he analyzed SNPs where the French, Chinese, and Papuans differed among each other, the values varied between 49.8 percent and 50.6 percent. So in all comparisons between people that did not involve Africans, the values were around 50 percent. But whenever an African and a non-African were compared, the Neanderthal matched the non-African at around 2 percent more SNPs than did the African. There did indeed seem to be a small but clearly discernible genetic contribution from Neanderthals to people outside Africa, no matter where they lived.
I read the two e-mails once. Then I read them again, this time very carefully, trying to catch any hint of a flaw in the analyses. I could find none. I leaned back in my office chair and looked blankly at my very untidy desk where papers and notes from the past few years had accumulated in layer after layer. David and Nick’s results stared at me from the computer screen. This was not a technical error of some sort. Neanderthals had contributed DNA to people living today. It was amazingly cool. It was what I had dreamed of achieving for the last twenty-five years. We had hard evidence to answer a fundamental question debated for decades about human origins, and the answer was unexpected. By showing that not all of the genomic information in present-day humans traced back to recent ancestors in Africa, it contradicted the strict out-of-Africa hypothesis of which my mentor Allan Wilson had been one of the main architects. It contradicted what I myself had believed to be true. Neanderthals weren’t totally extinct. Their DNA lived on in people today.
Staring blankly at my desk I realized that our results were unexpected not only in that they contradicted the out-of-Africa hypothesis. They also didn’t support the common version of the multiregional hypothesis. Contrary to the predictions of this hypothesis, we didn’t see the Neanderthal genetic contribution only in Europe, where Neanderthals had lived. We saw it also in China and Papua New Guinea. How could this be? Absentmindedly, I started to clean my desk. Slowly at first but then with increasing energy, I tossed out debris from years-old projects. Dust whirled into the air from layers deep on my desk. I needed to start a new chapter. I needed a clean desk.
Doing domestic tasks sometimes helps me think and as I cleaned, I visualized modern humans as arrows on a map coming out of Africa and meeting Neanderthals in Europe. I could imagine them having babies with Neanderthals—babies who then became incorporated among the modern humans, but I struggled to see how their DNA came to East Asia. It was possible that subsequent migration among modern humans might have brought Neanderthal DNA to China, but it seemed we would then find less similarity on average between a Chinese person and Neanderthals than between a European person and Neanderthals. Then it dawned on me: my imaginary arrows showing modern humans coming out of Africa passed through the Middle East! This was of course the first place where modern human would have met Neanderthals. If those humans mixed with the Neanderthals and then went on to become the ancestors of all people outside Africa today, the result would be that everyone outside Africa would carry approximately the same amount of Neanderthal DNA (see Figure 18.1). This must be a possible scenario. But I knew from experience that my intuition could sometimes be very wrong. Fortunately, I also knew that people like Nick, David, and Monty, who tested ideas mathematically, would set me straight if it was.
Figure 18.1. An illustration of the idea that if the Neanderthals mixed with early modern humans leaving Africa, and these went on to populate the rest of the world outside Africa, they would carry Neanderthal DNA with them to regions where Neanderthals never existed. For example, about 2 percent of the DNA of people even in China comes from Neanderthals. Photo: Pääbo, MPI-EVA.
We discussed David and Nick’s findings at our Friday meetings and during intense weekly consortium phone meetings. While some of us were now convinced that Neanderthals had mixed with modern humans, others were still reluctant to believe it, even though they struggled to explain how David and Nick’s analysis could possibly be wrong. I realized that if it was that hard to get everyone in our consortium to believe these results, we would have an even harder time convincing the world, especially the many paleontologists who saw no evidence for interbreeding with Neanderthals in the fossil record. These included some of the most respected people in the field, like Chris Stringer at the Natural History Museum in London and Richard Klein at Stanford University in California. Although I thought these paleontologists stood for a properly cautious interpretation of the fossil record, it still seemed possible that they had been influenced by previous genetic results. Many groups, including ours, had shown that the big picture of genetic variation in people today was one in which genetic variation had come out of Africa rather recently. Our 1997 paper showing that Neanderthals had not contributed any mtDNA to people today had also had a big influence. Although some paleontologists, such as Milford Wolpoff at the University of Michigan and Erik Trinkaus at Washington University in St. Louis, saw evidence of mixing in fossils, and some geneticists had made attempts to point out gene variants that might have come from Neanderthals, such arguments weren’t compelling enough to sway the common opinion. Or at least they hadn’t struck me as compelling. There had simply never before been a need to invoke any Neanderthal contribution to explain the patterns of either morphological or genetic variation across the world today. Now this situation had changed. We could look directly at the Neanderthal genome. And we saw a contribution, albeit a small one.
Still, I suspected we would need more to convince the world of our results. Science is far from the objective and impartial search for incontrovertible truths that nonscientists might imagine. It is, in fact, a social endeavor where dominating personalities and disciples of often defunct yet influential scholars determine what is “common knowledge.” One way to undermine this aspect of common knowledge would be to do additional analyses of the Neanderthal genome, independent of the counting of SNP alleles that David and Nick had done. If such additional and independent lines of evidence also suggested gene flow from Neanderthals into modern humans, then the world at large would be easier to convince. Finding what other
analyses we could do became a constant theme in our weekly phone meetings.
Somewhat unexpectedly, a viable suggestion came from outside our consortium. At the Cold Spring Harbor meeting in May 2009, David had met with Rasmus Nielsen, a Danish population geneticist who had done his PhD with Monty Slatkin back in 1998. He was now a professor of population genetics at UC Berkeley. Rasmus told David that he and his postdoc Weiwei Zhai had searched present-day genomes for regions that showed greater variation outside Africa than inside Africa. Although certainly possible, such a pattern is generally unexpected, as initially small offshoots of larger populations generally contain just a subset of the variation found in the ancestral group. If any such regions were found, there could be many explanations, but one possibility interested us very much. Since Neanderthals had lived independently from modern human ancestors for a few hundred thousand years outside Africa, they must have accumulated genetic variants distinct from those in modern humans. If they had subsequently contributed segments of their genome to people outside Africa, then Rasmus’s approach might identify genomic regions where this had happened because in such regions there should be just that pattern of more variation outside Africa than inside Africa. Using our Neanderthal genome, we could now check whether at least some of these regions came from Neanderthals, since the non-African versions of Rasmus’s regions would then be close to our Neanderthal DNA sequence. In June 2009, I asked Rasmus and Weiwei to join the Neanderthal Genome Analysis Consortium.
