First AASI mtDNA genomes from Sri Lanka (2500 and 5500 BC)

The mitochondrial genomes of two Pre-historic Hunter Gatherers in Sri Lanka:

Sri Lanka is an island in the Indian Ocean connected by the sea routes of the Western and Eastern worlds. Although settlements of anatomically modern humans date back to 48,000 years, to date there is no genetic information on pre-historic individuals in Sri Lanka. We report here the first complete mitochondrial sequences for Mesolithic hunter-gatherers from two cave sites. The mitochondrial haplogroups of pre-historic individuals were M18a and M35a. Pre-historic mitochondrial lineage M18a was found at a low prevalence among Sinhalese, Sri Lankan Tamils, and Sri Lankan Indian Tamil in the Sri Lankan population, whereas M35a lineage was observed across all Sri Lankan populations with a comparatively higher frequency among the Sinhalese. Both haplogroups are Indian derived and observed in the South Asian region and rarely outside the region.

No idea why this comes out of Sri Lanka first, and not India (bigger country), but it is what it is.

43 thoughts on “First AASI mtDNA genomes from Sri Lanka (2500 and 5500 BC)”

    1. Narasimhan tweeted this to Fernando…”Very interesting work! How do the contammix estimates look like? Do you have a plot of damage rates vs binned read length? ALG being 345x coverage + map damage plot G>A liftoff + being deeply nested amongst modern mtDNA of ALG is surprising. super excited for autosomal results!”

      Their mtDNA being deeply nested amongst moderns is not surprising to me since these samples are not very old.

  1. No surprises here, re M38a and M18a
    Mesolithic hunter-gatherers from two cave sites.
    Would be nice to know which cave sites, specially if they were in the same location as Balangoda man, (earliest- 30,000 BP reliably dated record of anatomically modern humans in South Asia).
    M18a was found at a low prevalence among Sinhalese, Sri Lankan Tamils, and Sri Lankan Indian Tamil in the Sri Lankan population,
    Any thoughts as to why. Comparison with India?.

    No idea why this comes out of Sri Lanka first, and not India (bigger country), but it is what it is.
    Maybe because older and better preserved samples for SL.
    Hopefully we might even get YDNA of Balangoda Man too.

    https://en.wikipedia.org/wiki/Balangoda_Man

  2. 5500 BC date is good, both are likely well before non-AASI intrusion.

    >No idea why this comes out of Sri Lanka first, and not India (bigger country), but it is what it is.

    Come one, we all kind of know why.

  3. Since those samples are from way down South, they are definitely AASI. Hopefully we can get them on G25.

    1. According to a study published in 2021 using 16 X-chromosomal short tandem repeat markers (STRs), there was no genetic subdivision detected between Sinhalese, Moors and Sri Lankan Tamils while Indian Tamils were having a subtle but statistically significant difference.

      1. This is likely due to caste differences, the up country Tamils of Sri Lanka (of recent Indian origin) are mainly of Dalit background, and are overwhelmingly haplotype M (>70%). Expect higher frequency of west Eurasian mtDNA in the upper castes.

  4. I parsed the paper quickly. The samples are from Udupiyangalge and Alugalge cave sites. UPG is dated to 8680+-30 BP (roughly 6600 BCE). ALG is dated to 4630+- 30 BP (roughly 2500 BCE).

    This is the first time M18 & M35 is being found in the ancient DNA record of South Asia.

    Since the autosomal analysis is going on (I hope so!), I want to know the nomenclatural procedure. Since there is no central body that arbitrates cline names, what will happen if Iran_Neolithic is a component of the SL Mesolithic Hunter-Gatherer population? Will we still call it Iran_Neolithic?

    1. Thanks Ugra

      Alugalge* is in Intermediate Zone and Udupiyangalge in the Dry Zone (not all that dry), East of the central mountain range (both not on google maps) . (See Figure 1 of Somadeva, 2021; link below). The large reservoir directly south of Udupiyangalge is the Uda Walawe Reservoir, a convenient landmark.

      Balangoda where Balangoda man was found is pretty much the heart of the wettest of the wet zone, on the southern slopes of the mountain range. Its the start of the steep climb into the central highlands. Balangoda is one of the few passes into the central high lands. Can be defended by a few people and why the Kandyan Kingdom did not get conquered till 1815 and then only by traitorous betrayal.

      Balangoda, now essentially tea country and Sinharaja is also part of tropical rain forest belt that includes the Amazon and the Congo.

      *Alugalge = grey-stone-house =alu:grey gal:stone ge (as in gay)=house..

      See below for photos and cultural artifacts
      https://rajsomadeva.com/cultural-dynamism/

    2. There is no chance that nuclear DNA from those two kids (they were like 10 years old) will show Iranian N ancestry, the archaeology is quite clear on this. Sri Lanka has no traces of a proper neolithic.

    3. How would we even infer Iran_N ancestry? Modeling each other of those Sri Lankans with Iran_N as one input + the other Sri Lankan as the second input is one way to see if one has more affinity than the other. The older one is predicted to have low coverage while the newer one is higher coverage.

      1. The same way we infer ancient Steppes contribution to modern Indians? Mesolithic is ancient to the Neolithic.

        1. The steppe contribution is measured with Indians as the output target, and the steppe component as the input (along with other inputs as well).

          To determine Iranian ancestry in these 2 samples what will be your inputs? Iranian + what else to figure if these have excess Iranian ancestry in comparison to that something else. The one method I proposed was to use each other as inputs like start with:

          1- Older Sri Lankan + Iranian as inputs, and the more recent Sri Lankan as the target. Here you could use various Iranians; Ganj Dareh, Hotu .etc

          2- Flip it, Younger Sri Lankan + Iranian as inputs, and the older Sri Lankan as the target. This is anachronistic however it could still be used to see if the older one has more Iranian affinity than the younger Sri Lankan.
          I am not saying anything new in this post, this is simply expounding what I said earlier about switching them around as inputs and outputs.

          But like what if both come up as 0% Iranian, someone could still say “well maybe they both have equal amounts so there is no excess, but it is undetected and actually there”. How can that be falsified? We’d have to wait for even older samples for that, or do you have another test in mind for a falsification?

          1. The modern Sri Lankan population is IE speaking. And parts of it are Dravidian speaking. In terms of a direct comparison, that will be the first target.

          2. How will modern populations be used to ascertain if ancient samples are pure AASI? I mean unless the ancient samples are far enough from the modern samples on an AASI cline, we’d need some way of falsification of the non-trivial presence hypothesis.

            One other way I have thought of is comparing Narasimhan’s constructs with these Southern AASI, but maybe the difference in the northern-derived AASI constructs vs southern actual AASI samples could possibly make some weird illusions as is the case with simulated populations every now and then.

      2. I don’t even understand what your question is, but there’s a bunch of tools pop geneticists use to figure out rough ancestry before even using formal statistics to model a sample. You can plot on all the samples on a PCA, do f3 outgroup runs, f4 stats, find out allele sharing between moderns and ancients, it’s not hard to see if those samples share alleles with IranN or not.

    1. Pure AASI could be an amalgamation of different proto-AASI. Maybe with slight ‘first’ Iran_N migrators input as well?

  5. According to Niraj Rai on Twitter, they are currently sequencing their nuclear DNA. The younger sample has good coverage; the older one is ok.

  6. So these are the people who represent the AASI component? Interesting! I hope they release the raw data or something because it would be interesting to see their scores on G25 or Harappaworld

  7. There is some news every now and then from Neeraj Rai about Ganga Mesolithic samples. I wonder what their uniparentals and autosomal profiles will be like. Those are from deep in India… not just the NW quadrant. Also their dates are pretty wild, the oldest being 12,000-10,000 years old (Mahadaha/Sanar Nahar Rai) and the most recent being 5000 to 6000 years old (Damdama). Along with the samples it would be nice if they could re-evaluate the dates of the specific samples.

    1. @DaThang I’m still with the hypothesis you proposed a few years back. Levantine Cro magnon + proto-AASI

      1. I’m not sure the exact words I used, but I suspect that Ganga Mesolithic is AASI + Iran HG. But the Sri Lankan HGs are probably pure AASI. As for AASI, yes I loosely suspect AASI = Levantine HG + proto-AASI, whereby the Levantine HGs in question are possibly Levant Aurignacian or Ahmarian.

  8. Razib, how long do you think it will take us for us to know their ydna/paternal haplogroup and their autosmal breakdown (i.e __% AASI+___% Neolithic Iranian, etc)? Or just their raw data file.

    1. We are unlikely to find any Neolithic Iran ancestry there. Both sites are hunter-gatherer sites.

    1. G
      My family is Tamil from Jaffna Sri Lanka, and my mtDNA is M35a, maybe my ancestor

      DNA research is saying is that in India (and for Sri Lanka*) our mothers ancestry (mtDNA) is the same, but our paternal lines (Y-DNA) can be different. One paper found that 70% of India including 26 tribal populations carried the M mtDNA haplogroup. In the Harappa DNA project 50% of the few Sri Lankan participants (8) had M mtDNA. Two of the four self identified Tamils and two of the four self identified Sinhalese.

      mtDNA lineages unite South Asians, while the Y lineages separate them (by caste and region). The generality has many exceptions, but it points to a peculiar sex mediated admixture process

      Macrohaplogroup M is ubiquitous in India and covers more than 70 per cent of the Indian mtDNA lineages The lineages M2, M3, M4, M5, M6, M18 and M25 are exclusive to South Asia, with M2 reported to be the oldest lineage on the Indian sub-continent. (Chandrasekar et al, 2009)

      See link below for references.

      https://sbarrkum.blogspot.com/2013/05/vijaya-kuveni-paradigm-for-m-mtdna-in.html

      1. @sbarrkum
        Some of that is very outdated information. mtDNA M appears to have founder effect in the region but M, R, N mtdna diversity is certainly local because Australian aboriginals derive from all three of those mtDNA. Only some clads of mtdna U might be from western eurasia but I’m not too attached to this idea for various reasons.

        They have also found root mtDNA M* in Andaman_100BP sample from Mayar et al. 2018 study. Which suggests there is some loss of M diversity on mainland due to founder effect.

        As far as Y-DNA goes – there is even bigger founder effect and bottelnecks. This is typical for Eurasia since Neolithic-Bronze Age from what we are seeing in published studies since 2016.

  9. @VNA
    Note that M haplogroup is root. Sub groups like M35 or M42 are because of mutations over thousands of years.

    No surprises but amazing that the female line of the first wave out of Africa continues to the present day from South Asia all the way to Australia.

    The first wave were those who left Africa around 50,000 years ago. Some ended up in Australia. The common linkage and continuity is the mtDNA M that most South Asians (M35, M42b etc) and Australian Aborigine ( M42a ) have inherited.

    Haplogroup M42a is widely accepted as indigenous to Australia, with M42a distantly related to M42b, which is found in the South Indian tribal groups, but with a divergence time of at least 55 KYA.
    Nagle et al 2016.
    https://www.nature.com/articles/jhg2016147

    1. @sbarrkum

      I’m aware. I was pointing out how M had founder effect and bottlenecks.

      Indigenous Vedda of Sri Lanka carry just 17% mtdna M, rest of their mtdna is R and U due to founder effect/bottleneck.

      1. Indigenous Vedda of Sri Lanka carry just 17% mtdna M, rest of their mtdna is R and U due to founder effect/bottleneck.

        Yes, from Ranaweera et al 2014. Excerpt from table below.

        M Haplogroup
        Vedda 17%
        Up-country Sinhalese 41.67%
        Low-country Sri Lankan 42.5%
        SL Tamils 43.5
        Indian Tamils (in SL) 70.18%

        Indian Estate Tamils are representative of the 70% Indian mtDNA M haplogroup percentage.
        Sri Lankan Sinhalese and Tamils have less percentages (approx 40+%) of mtDNA M haplogroup,

        Sri Lankan Sinhalese and Tamils also have R mtDNA. Note, Australian Aborigines also have R mtDNA.
        R mtDNA among Sri Lankans
        Vedda 45.33%
        Up Country Sinhalese 16.67%
        Low country Sinhalese 25%
        SL Tamils 7.69%
        SL Indian Tamils 8.77%

        mtDNA of Sri Lankans
        https://sbarrkum.blogspot.com/2015/06/mtdna-of-sri-lankans.html

  10. Misleading analysis by the authors. They state a higher frequency of these specific M haplotypes “among the Sinhalese”, but that’s because they have analysed 5 times more Sinhalese (12 out of 355). The actual prevalence is higher among both sets of Tamils (3 out of 69). See supplementary data.

Comments are closed.

Brown Pundits