Drivers of selection for ghosts in the genome

A new preprint on bioRxiv, Strong selective sweeps before 45,000BP displaced archaic admixture across the human X chromosome, is suggestive of an exciting new phase in human evolutionary genomics. Basically, leveraging whole-genomes in diverse populations to explore selection dynamics.

The authors looked at X chromosomes in males for reasons of technical tractability. Human males carry a single X chromosome, so it’s easy to determine the sequence of genetic variants across a physical stretch of DNA (females, with two X chromosomes, require phasing). The X is interesting for two other reasons: it is present in females about 2/3 of the time (because they have two copies), and, is subject to really strong selection in males 1/3 of the time. Basically, males exhibit no recessive dynamics on the X chromosome because we carry only one. This means that genetic variants which are “recessive” in their expression to selection in females are expressed in males.

The fact that the X is disproportionately found in females also means that all sorts of intra-genomic conflict driven by sex occur on this chromosome. You will know this if you read Matt Ridley’s excellent The Red Queen.

The specific result here is that the authors found a common family of haplotypes in males on the X chromosome outside of Africa whose homogeneity is indicative of a very strong sweep. From the text:

The identified selective sweeps are as strong or even stronger than the most dramatic sweeps previously found in humans. Ten sweeps span between 500kb and 1.8Mb in more than 50% of non-Africans (Table S2). The strongest sweep span 900kb in 91% of non-Africans and affects 53% of non-Africans across a 1.8Mb region. For comparison, the strongest sweep previously reported surrounds the lactase gene and spans 800kb in 77% of European Americans (24). The selection coefficient on the genetic variant driving this sweep was estimated to 0.15 (24) suggesting even stronger selection for several of the X chromosome sweeps we have identified.

The swept regions we identify here may be recurrent targets of strong selection during human evolution. To investigate this possibility, we intersect our findings with our previously reported evidence of selective sweeps in the human-chimpanzee ancestor (16).We find a strong overlap between the sweeps reported here and regions swept during the 2-4 my that separated the human-chimpanzee and human-gorilla speciation events (17, 25) shown as grey regions in Figure 2 (Jaccard stat.: 0.17, p-value: <1e-5) (Materials and Methods). This suggests that the identified regions of the X chromosome are continually subjected to extreme positive selection.

A selection coefficient of 0.15 is eye-popping. Selection coefficients of 0.01 are reasonable. For humans, anything in the 0.10 range is more like a weird artifact than a true result. But here it is. The fact that the regions overlap with earlier targets of selection during the speciation event that led up to our lineage is clearly of interest.

Looking more closely at the regions of the X which was subject to the selection, they found almost no archaic ancestry. That is, the ~1% Neanderthal ancestry that is expected across the X chromosome is almost absent in these segments derived from the selection event. The inference is made that perhaps then these sweeps occurred due to introgression from a sister modern human lineage, perhaps an earlier wave out of Africa which never mixed with Neanderthals. The archaeology is compelling now that these people existed, and there are tentative suggestions from genomics which attests to their presence as well (e.g., modern human admixture into the Altai Neanderthals).

Looking at the 45,000-year-old Siberian genome the authors found the same signatures that they see in other non-Africans. This means the event had to happen between 55 and 45 thousand years ago, after the Neanderthal admixture (which is found all around these zones in the genome), but before geographical diversification and expansion of the modern human lineage.

The authors conclude:

We hypothesize that our observations are due to meiotic drive in the form of an inter-chromosomal conflict between the X and the Y chromosomes for transmission to the next generation. If an averagely even transmission in meiosis is maintained by a dynamic equilibrium of antagonizing drivers on X and Y, it is possible that the main bottlenecked out-of-Africa population was invaded by drivers retained in earlier out-of-Africa populations. If this hypothesis is true, the swept regions represent the only remaining haplotypes from such early populations not admixed with Neanderthals.

Meiotic drive is a segregation distorter. A form of intra-genomic selection which is potentially very powerful. Some hypothesize that alleles normally subject to meiotic drive sweep through the population so fast that researchers underestimate the phenomenon’s ubiquity because they haven’t caught sweeps in action.

This strange evolutionary genetic process then may have preserved a genetic relic within the human genome. But on Twitter Iosif Lazaridis suggests that perhaps the donor population were “Basal Eurasians,” and all non-Africans may have some Basal Eurasian ancestry, with Near Easterners exhibiting more than baseline ancestry (presumably through later admixture).

A Kimura Age to the Kern-Hahn Era: neutrality & selection

I’m pretty jaded about a lot of journalism, mostly due to the incentives in the industry driven by consumers and clicks. But Quanta Magazine has a really good piece out, Theorists Debate How ‘Neutral’ Evolution Really Is. It hits all the right notes (you can listen to one of the researchers quoted, Matt Hahn, on an episode of my podcast from last spring).

As someone who is old enough to remember reading about the ‘controversy’ more than 20 years ago, it’s interesting to see how things have changed and how they haven’t. We have so much more data today, so the arguments are really concrete and substantive, instead of shadow-boxing with strawmen. And yet still so much of the disagreement seems to hinge on semantic shadings and understandings even now.

But, as Richard McElreath suggested on Twitter part of the issue is that ultimately Neutral Theory might not even be wrong. It simply tries to shoehorn too many different things into a simple and seductively elegant null model when real biology is probably more complicated than that. With more data (well, exponentially more data) and computational power biologists don’t need to collapse all the complexity of evolutionary process across the tree of life into one general model, so they aren’t.

Let me finish with a quote from Ambrose, Bishop of Milan, commenting on the suffocation of the Classical religious rites of Late Antiquity:

It is undoubtedly true that no age is too late to learn. Let that old age blush which cannot amend itself. Not the old age of years is worthy of praise but that of character. There is no shame in passing to better things.

A historical slice of evolutionary genetics

A few friends pointed out that I likely garbled my attribution of who were the guiding forces between the “classical” and “balance” in the post below (Muller & Dobzhansky as opposed to Fisher & Wright as I said). I’ll probably do some reading and update the post shortly…but it did make me reflect that in the hurry to keep up on the current literature it is easy to lose historical perspective and muddle what one had learned.

Of course on some level science is not as dependent on history as many other disciplines. The history is “baked-into-the-cake.” This is clear when you read The Origin of Species. But if you are interested in a historical and sociological perspective on science, with a heavy dose of narrative biography, I highly recommend Ullica Segerstrale’s Defenders of the Truth: The Battle for Science in the Sociobiology Debate and Beyond and Nature’s Oracle: The Life and Work of W.D. Hamilton.

Defenders of the Truth in particular paints a broad and vivid picture of a period in the 1960s and later into the 1970s when evolutionary thinkers began to grapple with ideas such as inclusive fitness. E. O. Wilson’s Sociobiology famously triggered a counter-reaction by some intellectuals (Wilson was also physically assaulted in the 1978 AAAS meeting). Characters such as Noam Chomsky make cameo appearances.

Segerstrale’s Nature’s Oracle focuses particularly on the life and times of W. D. Hamilton, though if you want that at high speed and max density, read Narrow Roads of Gene Land, Volume 2. Because Hamilton died before the editing phase, the biographical text is relatively unexpurgated. Hamilton also makes an appearance in The Price of Altruism: George Price and the Search for the Origins of Kindness.

The death of L. L. Cavalli-Sforza reminds us that the last of the students of the first generation of population geneticists are now passing on. With that, a great of history is going to be inaccessible. The same is not yet true of the acolytes of W. D. Hamilton, John Maynard Smith, or Robert Trivers.

W. D. Hamilton, Darwin’s 20th century heir

Today on Twitter there was a discussion about why there wasn’t a biography of John Manyard Smith. One reason might be that John Maynard Smith was a pretty nice and congenial fellow. There wasn’t much excitement from what I know.

In contrast, if you read R.A. Fisher: The Life of a Scientist, you get the sense that he was a bit of a dick (the book was written by his daughter). Of course, Fisher was a great scientist, an eminence is both statistics and evolutionary biology. Nevertheless, his irascible personality lends itself to biographical treatment, though rarely hagiographical (a friend is writing a book on Fisher’s life; he too confirms, the guy was a dick).

The evolutionary biologist who yearned to be Fisher’s heir, W. D. Hamilton, was not a dick, but something of a dope. Hamilton has had a full-length biography devoted to him, Nature’s Oracle. But he makes appearances in Defenders of the Truth: The Sociobiology Debate, The Price of Altruism: George Price and the Search for the Origins of Kindness, and A Reason for Everything: Natural Selection and the British Imagination.

I say Hamilton was a dope because he was socially awkward, and obviously got himself in trouble through his guilelessness. If he were alive today Hamilton would be in a whole lot of trouble I suspect, except for the fact that he’d be emeritus. Unfortunately, Bill Hamilton died in 2000 due to malaria contracted from field work (or malaria medication).

Yet Hamilton has left us two books where provides autobiographical sketches interleaved between his scientific papers, Narrow Roads of Gene Land: Evolution of Social Behavior, and Narrow Roads of Gene Land: Evolution of Sex (the third volume has biographical sketches from collaborators). In case you are not aware, Hamilton was the originator of the idea of “inclusive fitness.” At the same time, John Maynard Smith was developing “kin selection.” Hamilton distrusted Maynard Smith because of this coincidence, suspecting some sort of scientific fraud (both of them were in communication with George Price at the time).

Narrow Roads of Gene Land: Evolution of Sex was published without revisions to a very long draft of autobiographical sketches because Hamilton had died. It is quite a rambling and sometimes incoherent piece of work because editors couldn’t give any feedback. But it’s fascinating because it’s an unvarnished window into Hamilton’s strange brain.

Of course, the primary reasons to read the three volumes on the scientific papers. I’ve read the famously notationally-inscrutable paper on inclusive fitness published in 1964 many a time. Bill Hamilton had an interesting life and a quirky mind. I’m quite sad that he’s not here anymore.

Idle theories are the devil’s workshop

In the 1970s Richard C. Lewontin wrote about how the allozyme era finally allowed for the testing of theories which had long been perfected and refined but lay unused like elegant machines without a task. Almost immediately the empirical revolution that Lewontin began in the 1960s kickstarted debates about the nature of selection and neutrality on the molecular level, now that molecular variation was something they could actually explore.

This led to further debates between “neutralists” and “selectionists.” Sometimes the debates were quite acrimonious and personal. The most prominent neutralist, Motoo Kimura, took deep offense to the scientific criticisms of the theoretical population geneticist John Gillespie. The arguments around neutral theory in the 1970s eventually spilled over into other areas of evolutionary biology, and prominent public scientists such as Richard Dawkins and Stephen Jay Gould got pulled into it (neither of these two were population geneticists or molecular evolutionists, so one wonders what they truly added besides bluster and publicity).

Today we do not have these sorts of arguments from what I can tell. Why? I think it is the same reason that is the central thesis of Benjamin Friedman’s The Moral Consequences of Economic Growth. In it, the author argues that liberalism, broadly construed, flourishes in an environment of economic growth and prosperity. As the pie gets bigger zero-sum conflicts are attenuated.

What’s happened in empirical studies of evolutionary biology over the last decade or so is that in genetics a surfeit of genomic data has swamped the field. Some scholars have even suggested that in evolutionary genomics we have way more data than can be analyzed or understood (in contrast to medical genomics, where more data is still useful and necessary). Scientists still have disagreements, but instead of bickering or posturing, they’ve been trying to dig out from the under the mountain of data.

It’s easy to be gracious to your peers when you’re rich in data….

The Loneliest Neanderthal

Neanderthals are in the news again! This is good for me personally, as my company is selling Neanderthal trait analysis. Ooga-booga!

In any case, the two papers which have triggered the current wave of Neandermania are The Contribution of Neanderthals to Phenotypic Variation in Modern Humans, and A high-coverage Neandertal genome from Vindija Cave in Croatia. They are somewhat different. The first publication looks at introgressed archaic variants within modern populations. The second gets some results out of a much higher quality European Neanderthal which lived ~50,000 years ago.

The cool thing about the first paper is that it combined UK Biobank data, 100,000+ individuals, with hundreds of thousands of markers, and Neanderthal genomic data. Note that: a paper comparing ancient genomes with over 100,000 individuals and hundreds of thousands of markers. Now that’s 2017!

To find archaic alleles they:

  1. Looked for variants fixed in Yoruba (no Neanderthal), and homozygote or heterozygote in the alternative state in the Altai Neanderthal, which also segregated (varied) in the UK Biobank population. Basically, an allele not found in Africans but found in Neanderthals, and also found in appreciable fractions in the UK Biobank data set.
  2. They then took the SNPs above, and only retained ones confidently embedded in tracts of Neanderthal ancestry. Haplotype was consistent with admixture ~50,000 years ago (the length), and exhibited lower distance to Neanderthal than African genomes.

They did some stuff with tag-SNPs though. Overall they found a lot of the usual suspects. Pigmentation. Chronotype. But this passage jumped out at me:

In fact, for most associations, Neanderthal variants do not seem to contribute more than non-archaic variants. However, there are four phenotypes, all behavioral, to which Neanderthal alleles contribute more phenotypic variation than non-archaic alleles: chronotype, loneliness or isolation, frequency of unenthusiasm or disinterest in the last 2 weeks, and smoking status.

What they are saying is that for a lot of traits Neanderthals don’t really change the direction of the trait in humans, they just add more variants. This seems to be the case in pigmentation. Entirely unsurprising, Neanderthals were around for hundreds of thousands of years. Of course they had a lot of variation amongst themselves.

But the behavioral traits above shifted the modern humans in the aggregate who had the archaic allele somewhat. That is, being Neanderthal derived made a difference.

There have long been speculations about the sociality (or lack thereof) of Neanderthals. It would not be surprising if small population sizes meant that Neanderthals were less gregarious than modern humans, and that their lack of gregariousness did not redound to their benefit when they encountered the last wave of moderns.

Which brings us to the second paper. The big deal here is that it gives us a very high quality ancient genome of a European Neanderthal that lived ~50,000 years ago (the Vindija sample). Before this we had a high quality ancient genome of an Asian Neanderthal that lived ~125,000 years ago (Altai sample).  ~75,000 years is a long time. It’s so long that almost all the ancestry of modern non-Africans would have converged to a common population that long ago. Additionally, all the available data indicate that most of the admixture into modern humans from Neanderthals occurred around 50,000 years ago. So this new sample is definitely welcome.

It is not surprising that the Vindijia sample seems to be closer to the Neanderthal admixture population than the Altai sample. First, it is likely geographically closer, since all non-African populations have some Neanderthal ancestry West Asia is probably the top candidate, and southeastern Europe is not that far from West Asia in comparison to Mongolia. Second, it is basically contemporaneous with the Neanderthals who contributed ancestry to modern humans who left Africa. This means that the Neanderthal admixture percentage in non-Africans goes up moderately.

To me this is the most important paragraph:

It has been suggested that Denisovans received gene flow from a human lineage that diverged prior to the common ancestor of modern humans, Neandertals and Denisovans (2). In addition, it has been suggested that the ancestors of the Altai Neandertal received gene flow from early modern humans that may not have affected the ancestors of European Neandertals (13). In agreement with these studies, we find that the Denisovan genome carries fewer derived alleles that are fixed in Africans, and thus tend to be older, than the Altai Neandertal genome while the Altai genome carries more derived alleles that are of lower frequency in Africa, and thus younger, than the Denisovan genome (20). However, the Vindija and Altai genomes do not differ significantly in the sharing of derived alleles with Africans indicating that they may not differ with respect to their putative interactions with early modern humans (Fig. 3A & B). Thus, in contrast to earlier analyses of chromosome 21 data for the European Neandertals (13), analyses of the full genomes suggest that the putative early modern human gene flow into Neandertals occurred prior to the divergence of the populations ancestral to the Vindija and Altai Neandertals ~130-145 thousand years ago (Fig. 2). Coalescent simulations show that a model with only gene flow from a deeply diverged hominin into Denisovan ancestors explains the data better than one with only gene flow from early modern humans into Neandertal ancestors, but that a model involving both gene flows explains the data even better. It is likely that gene flow occurred between many or even most hominin groups in the late Pleistocene and that more such events will be detected as more ancient genomes of high quality become available.

These results seem to support earlier work indicate that Denisovans were admixed with an ancient hominin group which diverged very early on (probably the descendents of East Asia erectus?). And, that Neanderthals received gene flow from a lineage of modern (African?) humans 150,000 or more years ago. Since the latest work suggests that modern humans in some form have existed between from 200,000 to 350,000 years ago, this is entirely plausible.

But, it brings us the take-home message that the emergence of Pleistocene humanity was to a some extent characterized by reticulate gene flow, rather than a bifurcating tree.