April 14, 2014

IBD sharing between modern humans, Denisovans and Neandertals

bioRxiv doi:doi: 10.1101/003988

Sharing of Very Short IBD Segments between Humans, Neandertals, and Denisovans

Gundula Povysil, Sepp Hochreiter

We analyze the sharing of very short identity by descent (IBD) segments between humans, Neandertals, and Denisovans to gain new insights into their demographic history. Short IBD segments convey information about events far back in time because the shorter IBD segments are, the older they are assumed to be. The identification of short IBD segments becomes possible through next generation sequencing (NGS), which offers high variant density and reports variants of all frequencies. However, only recently HapFABIA has been proposed as the first method for detecting very short IBD segments in NGS data. HapFABIA utilizes rare variants to identify IBD segments with a low false discovery rate. We applied HapFABIA to the 1000 Genomes Project whole genome sequencing data to identify IBD segments which are shared within and between populations. Some IBD segments are shared with the reconstructed ancestral genome of humans and other primates. These segments are tagged by rare variants, consequently some rare variants have to be very old. Other IBD segments are also old since they are shared with Neandertals or Denisovans, which explains their shorter lengths compared to segments that are not shared with these ancient genomes. The Denisova genome most prominently matched IBD segments that are shared by Asians. Many of these segments were found exclusively in Asians and they are longer than segments shared between other continental populations and the Denisova genome. Therefore, we could confirm an introgression from Deniosvans into ancestors of Asians after their migration out of Africa. While Neandertal-matching IBD segments are most often shared by Asians, Europeans share a considerably higher percentage of IBD segments with Neandertals compared to other populations, too. Again, many of these Neandertal-matching IBD segments are found exclusively in Asians, whereas Neandertal-matching IBD segments that are shared by Europeans are often found in other populations, too. Neandertal-matching IBD segments that are shared by Asians or Europeans are longer than those observed in Africans. This hints at a gene flow from Neandertals into ancestors of Asians and Europeans after they left Africa. Interestingly, many Neandertal- or Denisova-matching IBD segments are predominantly observed in Africans - some of them even exclusively. IBD segments shared between Africans and Neandertals or Denisovans are strikingly short, therefore we assume that they are very old. This may indicate that these segments stem from ancestors of humans, Neandertals, and Denisovans and have survived in Africans.

Link

13 comments:

Rokus said...

This focus on the current distribution of purported Neanderthal and Denisovan IBD segments is a valuable extenstion to previous previous results described in Hochreiter's "Rare Haplotypes in the Korean Population" (2012) and "HapFABIA: Identification of very short segments of identity by descent characterized by rare variants in large sequencing data" (2013) that still may not have received due attention or assessment yet. Especially the inflated portion of European-African shared DNA snippets in the African-European-Asian triangle - despite the apparent higher archaic contribution in Asians! - that already caught my attention in Expanding Hybrids And The Rise Of Our Genetic Common Denominator, remains without due attention
I found the accumulated European contribution as compared to the Asian contribution to be even more extreme in this new paper (8,6% versus 4,6%, apparently due to the exclusion of Asian private variation) than in mentioned previous publications (2012: 8,4% versus 6,1%; 2013: 11,8% versus 6,8%). However, it goes without saying that Asians should have more archaic DNA than Europeans. Remarkably, this combines with massive attestation of European-African shared segments. Reluctance to subsequently consider this shared European-African portion as essentially European seems to be the logical consequence of embracing Out of Africa. However, the African tendency for smaller sized IBD segments being noted in the article, doesn't seem to involve the shared European-African portion in any significant way! Publication 2013 Figure 8 already revealed that the distribution of IBD segment sizes does not differ so much between Asian and European samples, as compared with on average shorter African samples - except for a second peak for the Asian sample that might indicate an additional, more recent Asian admixture event. This second "Asian" peak, though small, may indeed be the true cause of the Asian excess of archaic DNA, apparently without having this significantly increased the (still lower) total amount of Asian IBD segments.

Ryan said...

"Interestingly, many Neandertal- or Denisova-matching IBD segments are predominantly observed in Africans - some of them even exclusively."

Does this represent some sort of earlier back-migration of Neanderthal or Denisovian-like archaics into Africa? I wonder if these would be the same archaics that Hammer et al identified as having contributed to some of the populations in Africa (particularly hunter-gatherer groups IIRC).

If so that wouldn't be all that surprising I guess, but somewhat reassuring, as it fits the view that these groups were our peers and ancestors, and not simply our victims.

Tom Bridgeland said...

Ryan, I suspect this is just another example of Africa being more diverse in general, so they maintain ancestral segments that out of Africans don't. A lot of variation was lost when only a small population left Africa.

Mark Moore (Moderator) said...

1) I was surprised at the human IBD diversity in Amerindians. They (Mexicans I think) had far more than EUR or ASN groups and second only to AFR.

2) Africans shared a lot of IBDs with Neandertals, but the authors said the short average length of the segments probably means this is an artifact of the time they were all one population. Maybe one of you pros can explain why this is not at odds with the study above this one which says Eurasians are different because of admixture, not population substructure in Africa. It seems to me if there are that many more Archaic IBDs floating around Africa than elsewhere, then it could be substructure. The AVERAGE length may have been shorter, but still IBDs of various lengths were found in AFR.

3) Intuitively I would think that the shorter the IBD, the more likely it is that it is not an IBD at all, but they look alike by chance. Especially in regions where 95% of the gene has to be alike to get a functioning gene.

4) I wonder if some of the difference in average IBD length might be explained by different mutation rates or different generation times. If one group had an average generation time of 20 years and the other 30 years, one might expect shorter IBDs on average in the group which has a shorter generation time due to their having more generations, not their being further back in time.

Ryan said...

Fair point Tom. I guess this method wouldn't be able to say anything one way or another about older admixture events would it? Pity.

Rokus said...

@Mark,
'I was surprised at the human IBD diversity in Amerindians. They (Mexicans I think) had far more than EUR or ASN groups and second only to AFR.'

Not a surprise at all, since the samples were not Amerindian but Admixed Americans. Their high diversity must be due to the portion of African ancestry.

'Africans shared a lot of IBDs with Neandertals, but the authors said the short average length of the segments probably means this is an artifact of the time they were all one population.'

First of all, this effect is not noticeable for the shared European-African part. This part also contributes more to the total amount of Neanderthal DNA for containing most of the larger segments.
The overwheling majority of archaic African segments, however, is private African and clusters with Neanderthal and/or Denisovan in one degree or the other.

The private-African Denisovan (7,000 bp) and Neanderthal (6,000 bp) curves and the private-European Denisovan curve (12,000 bp) all appear to be older and to refer to unrelated admixture events. It can't be excluded these events included hominines that were actually less related to the Neanderthal and Denisovan reference samples.

'Intuitively I would think that the shorter the IBD, the more likely it is that it is not an IBD at all, but they look alike by chance.'

The IBD segments all have some rare mutations that are unlikely to have occurred independently.
But of course, there must be a limit in breaking up segments by cross-over before the IBD simply disappear below detection level. Or maybe these fragmented, unrecognizable pieces just accumulated again within the most common segments.

'I wonder if some of the difference in average IBD length might be explained by different mutation rates or different generation times.'

Indeed, there are many different averages of superimposed distribution curves that still have to be decomposed into its constituent admixture events. Most relevant for evaluating the different spikes may be the effects of natural selection.

eurologist said...

Mark,

@ (3) The authors surely excluded regions that were identical in all compared populations.

@ (4) Since mutations mostly derive from the male lineage, and there they accumulate roughly linearly with the age of the person, generation length does not enter at all - only the amount of mutations over time.

Unknown said...

Thank you for all your comments. We really appreciate the discussion. We will try to answer some open questions and add new ones that we were not able to answer.

@ Mark:

(2) The study of Lohse and Frantz that you mentioned looked at admixture from Neandertals into Eurasian populations after their expansion out of Africa to explain the higher rates of sharing between the Neandertal genome and Eurasians compared to Africans. In our analysis this corresponds to the high number of relatively long IBD segments that Europeans and Asians share with the Neandertal genome, which we also assume to be due to admixture events after their ancestors left Africa because the segments are rather long.

The surprisingly large number of short IBD segments shared between Africans and Neandertals in contrast could be due to substructure within Africa before ancestors of Eurasians moved out of Africa. They may be remnants of a common ancestor of Neandertals and humans or due to very early admixture events. Because of population substructure at this time it might be, that these IBD segments only survived in groups that stayed in Africa and are therefore exclusively found in present day Africans.

(3) As eurologist and Rokus already pointed out, our detected IBD segments are tagged by rare variants (tagSNVs), therefore it is not likely that many individuals share a lot of these rare variants by chance. The false discovery rate is governed by the number of tagSNVs. Short segments have often many tagSNVs: e.g. 3kbp segments have 50 SNVs. We found that short segments are tagged as good as long segments, therefore the false discovery rate should not increase for short segments.


Questions that came up and that we could not answer:

I) What we do not understand: **Where do the tagSNVs come from?** There are accumulations of these rare tagSNVs locally on the chromosome. Do they come from genomes that evolved separately and are subsequently introduced into the genome of ancestors of humans? Do they stem from mutation events that introduce many mutations at once?

II) Another strange fact can be seen in Figure 10. There are more Archaic segments than Denisovan segments. We defined Archaic as the intersection of Neandertal and Denisova, that is, IBD segments that match both the Neandertal and the Denisova genome. If an IBD segments matches the Denisovan genome, then in more than half of the cases it also matches the Neandertal genome. Figure 9 tells that the hybrid (Archaic) genome is more prominent in Asia. Where do these segments come from?


Gundula Povysil and Sepp Hochreiter

Mark Moore (Moderator) said...

Thank you doctors Povysil and Hochreiter, and Rokus and eurologist, for your answers. There is a lot to chew on.

Dienekes, here is one that I thought for sure you would post so that the brain trust could delve into it, but have not seen it up yet... http://www.sciencedaily.com/releases/2014/04/140421164242.htm

eurologist said...

"Another strange fact can be seen in Figure 10. There are more Archaic segments than Denisovan segments. We defined Archaic as the intersection of Neandertal and Denisova, that is, IBD segments that match both the Neandertal and the Denisova genome. If an IBD segments matches the Denisovan genome, then in more than half of the cases it also matches the Neandertal genome. Figure 9 tells that the hybrid (Archaic) genome is more prominent in Asia. Where do these segments come from?"

As I have said in the past, I think there are primarily two reasons for this. Firstly, I believe West Eurasian heidelbergensis decoupled from African heidelbergensis around ~350,00 - 400,000 ya, and much of Neanderthal is derived from that (but not all, as the more recent and more African/ AMH-like mtDNA of Neanderthals tells us).

So, we would expect some strong but not total correlation between Eurasian heidelbergensis and Neanderthal.

On the flip side, both Denisovans and modern Asians likely have additional admixture that introgressed from Asian erectus into eastern heidelbergensis before eastern admixture with AMHs.

Ryan said...

"Another strange fact can be seen in Figure 10. There are more Archaic segments than Denisovan segments. We defined Archaic as the intersection of Neandertal and Denisova, that is, IBD segments that match both the Neandertal and the Denisova genome. If an IBD segments matches the Denisovan genome, then in more than half of the cases it also matches the Neandertal genome. Figure 9 tells that the hybrid (Archaic) genome is more prominent in Asia. Where do these segments come from?"

I think this suggests that at the very least, the barriers between various archaic human populations were not particularly large, and that admixture may have been a common and continuous process for multiple groups over a very long time. Does that sound about right?

Ryan said...

It would also support John Hawkes' suggestion that Neanderthals originated from an earlier Out-of-Africa expansion that admixed with a relative of Denisovians.

eurologist said...

It would also support John Hawkes' suggestion that Neanderthals originated from an earlier Out-of-Africa expansion that admixed with a relative of Denisovians.

Ryan,

I have not read John Hawks' comments on this, but surely the researchers to credit for this idea are the ones who determined that Neanderthal mtDNA is more closely related to AMHs than heidelbergensis' (Denisovan) mtDNA is - see also my comment above.