Substack cometh, and lo it is good. (Pricing)

Selection and the Neanderthal Genome

For about ten years now there have been debates and discussions about the nature of Neanderthal admixture across Eurasia. How many admixture pulses were there? What is the relationship between admixture and selection? The fact is that West Eurasians have less Neanderthal admixture than East Eurasians. One explanation is that there was further admixture in East Eurasians. Another is dilution of Neanderthal admixture in West Eurasians due to later migration from Africans or Basal Eurasians (both of which lack Neanderthal admixture). But it could also be that the power of selection varies across Eurasia, with the lower population size of East Eurasians over the long-term resulting in less removal of Neanderthal ancestry from their genomes.

I think the main answer to the variation has to do with dilution. In particular, gene-flow from a very Basal Eurasian group into West Eurasians.

But, that doesn’t mean I don’t think selection has had an impact. A new preprint gets at this, Quantifying the contribution of Neanderthal introgression to the heritability of complex traits:

…We integrate recent maps of Neanderthal ancestry with well-powered association studies for more than 400 diverse traits to estimate heritability enrichment patterns in regions of the human genome tolerant of Neanderthal ancestry and in introgressed Neanderthal variants themselves. First, we find that variants in regions tolerant of Neanderthal ancestry are depleted of heritability for all traits considered, except skin and hair-related traits. Second, the introgressed variants remaining in modern Europeans are depleted of heritability for most traits; however, we discover that they are enriched for heritability of several traits with potential relevance to human adaptation to non-African environments, including hair and skin traits, autoimmunity, chronotype, bone density, lung capacity, and menopause age. To better understand the phenotypic consequences of these enrichments, we adapt recent methods to test for consistent directional effects of introgressed alleles, and we find directionality for several traits. Finally, we use a direction-of-effect-aware approach to highlight novel candidate introgressed variants that influence risk for disease…

The basic result seems to be that outside of a few characteristics, a lot of the Neanderthal variation is just not retained in the human genome. Traits like height, BMI, and EDU, are highly polygenic. So the genome-wide selection against Neanderthal alleles would be important.

10 thoughts on “Selection and the Neanderthal Genome

  1. Is there any degree to which this is quantified in the paper, in terms of contribution to genome wide ancestry? The conclusion that there is some turnover in Neanderthal variants is pretty plausible, but what seemed implausible to me is reduction in 50% N genome wide ancestry from it. Which is why the Petr paper suggesting no real change in genome-wide made sense to me (https://www.pnas.org/content/116/5/1639).

    Also: How much do comparable age AMH samples (50kya) have different modal variants than modern people at these N heritability depleted trait domains? Is heritability enrichment of AMH ancestry in these domains driven by variants that were frequent in ancient AMH, or variants which were minor alleles then and turned over later in prehistory?

  2. From the paper:
    “We first considered the loci most stronglycontributing to the genome-wide significant299positive correlationbetween Neanderthal LD profile and sunburn risk. One such windowwitha 300strongpositivecorrelation(chr9:16641651–16787775, R = +0.83) overlapped the gene BNC2301which influencesskin pigmentation levels in Europeans(Fig. S8).56As previously reported, 302BNC2has strong support for the association of introgressed variation with risk for sunburn: it is 303overlapped by a Neanderthal haplotype at ~70% frequency in Europeans associated with sun 304sensitivity17,57and is near archaic alleles that tag introgressed haplotypes (rs10962612tagging 305chr9:16,720,122–16,804,167; rs62543578 tagging chr9:16,891,561–16,915,874) strongly 306associated with increased incidence of childhood sunburn and poor tanning in the UK Biobank.4307We report 29other windows associated with sunburn using this methodin Table S6.Among 308these regions, we identifymany promising candidates, includingone nearby SPATA33 which 309has been implicated in tanning response, facial pigmentation, and skin cancer58and one nearby 310MC1Rwhich isa key genetic determinant of pigmentation and hair color.”

    Is there any confirmation of Neandertals having straight hair, sometimes non-black coloration? Especially red hair?

  3. But it could also be that the power of selection varies across Eurasia, with the lower population size of East Eurasians over the long-term resulting in less removal of Neanderthal ancestry from their genomes.

    Thinking about this idea some more… Using distance from Yoruba and the whole-genome sequences from ChinaMap and 1000 Genomes (https://www.nature.com/articles/s41422-020-0322-9), then taking Fst from Yoruba (Fig 4) as a proxy for smaller population size, because smaller population = more genetic drift = higher Fst, then GBR-YRI (Great Britain – Yoruba) is about 0.157 vs CHB-YRI (China Beijing – Yoruba) 0.168. So I don’t know, doesn’t seem that large.

    Some Fsts I have calculated on Human Origins data actually has a larger difference of English-Yoruba of 0.152 vs Han-Yoruba of 0.184. It seems like the whole-genome sequences reduce Fst between East Asian groups and other populations. E.g. Han-Tuscan is 0.111 on Human Origins data, only 0.098 on whole genome data.

    (No particular relevance of UK and China populations, they are just examples of West Eurasian and East Eurasian populations that are available.)

    If East Asian groups don’t seem to have a lot more drift overall, particularly on whole-genome sequences that would get around ascertainment(?), then how much is it plausible that they evolved in much smaller enough population sizes anyway to reduce efficiency of selection against Neanderthal ancestry?

    Obviously this is quite a crude way of inferring about the population sizes over time, but I think it kind of suggests some more plausibility to the idea of dilution of Neanderthal ancestry.

  4. What ydna would you guys associate with Basal Eurasian?

    Also I read this interesting paper.

    https://www.biorxiv.org/content/10.1101/867317v1.full.pdf

    Would love to see ydna > 45,000 years old.

    ———————————————————–‘
    Also is there any more Ancient DNA coming out from North Eurasia? I’m curious to see if my hypothesis about y C being the West Eurasian side of ANE is true. I’m also curious which ENA populations admixed into ANE. For example is Tianyuan directly ancestral to Yana and is Yana directly ancestral to Malta?

  5. @ Jatt

    One thing that’s interesting regarding Eurasian uniparentals is mtDNA – the primary branches of both N and M are mostly East Eurasian (especially M, which has like 40 or 50 primary branches alone, all of them clearly eastern except M1, and quite a few of these branches are obviously very deeply rooted in Asia). But I think we have a problem if we assume Eurasian Y or mtDNA spread from East to West, because we know autosomally that East Eurasians are noticeably more bottlenecked/inbred then West Eurasians. So how likely is it that we have a scenario where a few Eastern lineages intrude westward, come in contact with a larger population, and replace the “native” Y’s of that numerically larger group? Granted, we know that from the Neolithic onwards this wasn’t unusual, smaller groups with better technology or a more strongly militaristic social structure did this all the time, but we’re talking about rougly 50,000 kya hunter-gatherer groups here, where I would think the raw numbers of people would count for more.

  6. Also regarding Yana, if I recall correctly the ENA affinity he showed seemed to be basal to Tianyuan. Not sure how closely related he was to Malta, I’m sure it was covered in the original study though.

    I also find it impressive how many (nearly) basal branches of East Eurasians there are so far. Tianyuan’s his own branch, then you have Onge-Hoabhinian-AASI who I think form a very weak clade together (they all might share more drift with each other than any of them do separately with other East Eurasians, but they’re all still highly diverged from each other), and then Australians+Papuans, again also deeply diverged themselves, and finally this cryptic ENA branch that exists in Yana. Contrast this with West Eurasia where we seem to have a population turnover in Europe/Russia after 40,000 kya with the Gravettian, then a possible reflux of WHG-ish ancestry to the Near East after the LGM. More internal population churn in West Eurasia, it seems.

  7. @Mick

    we know autosomally that East Eurasians are noticeably more bottlenecked/inbred then West Eurasians

    Is not this because of the Basal Eurasian admixture in West Eurasians? Non-Basal Eurasian-admixed hunter-gatherers of Europe were also quite bottlenecked/inbred, if not more, as far as I know.

  8. @Mick, I haven’t really looked into the details of that paper yet, sorry.

    @Onur, like my above, just going by the Fst data on Human Origins and assuming that 1) Fst from African populatons = approximate proxy for population size / bottlenecks, and 2) that East Asian Fst is a little inflated in HumOr compared to Whole Genome, while ancient European capture dna would not be… it would seem that the late Euro HG populations would be a little more drifted / bottlenecked than most present-day East Eurasian populations, East Asians for example. Less sure about Papuans, Onge, Native Americans and even less about the other smaller populations that contributed less to present day peoples (Jomon, Hoabinhians).

    On the other hand, the “steppe” population and European EEF would both have an overall Fst from African pop about the same and slightly lower than present-day East Asian populations (then more pooling of their ancestry variation and other streams in, that affect mainly southern and northeastern Europe gets to where different European populations are today).

    Some of the populations which are rich in Basal Eurasian like Natufian *seem* to have high Fst from Africans as well, in the same kind of ballpark as late UP/Meso Euro HG. (Hence small, heavily drifted population.)

    I would guess that the influence of the Basal Eurasian ancestry could have resulted in higher effective autosomal population size in West Eurasians, but seems more like because pooling of variation, rather than because of Basal Eurasian populations continuously maintained literally a high population size themselves.

    It actually seems quite difficult to answer these questions on ancient dna basis using tools like Fst. Even though the dna is good for population genetics questions, I think ascertainment might affect the ability to draw out these inferences about population size and drift a little bit, and then compare slightly different present day data and ancient data…

Comments are closed.