Thursday, July 21, 2022

The Testimony Of 14,000 Year Old Ancient DNA From SW China

In Xiaoming Zhang et al., "A Late Pleistocene human genome from Southwest China" (2022), the authors discuss their analysis of 14,000 year old ancient DNA from a woman at the Red Deer Cave named Mengzi Ren (MZR). (I tip my hat to Bernard's blog and to Razib who together alerted me to this paper.) I previously blogged this find, before ancient DNA analysis was available, on December 18, 2015.

Her "mosaic" of seemingly partially archaic bone features suggested that she might be an archaic hominin-modern human hybrid in Southwest China. 

As was the case with seemingly archaic featured remains in South America that were later genotyped to be ordinary indigenous Americans genetically, the reality was less flashy and was largely paradigm affirming. 

The results disfavor the late persistence of archaic hominins in mainland Asia, and demand a non-genetic explanation for her seemingly archaic features, as she is genetically a typical Southeast and southern East Asian anatomically modern human.

Perhaps the most novel conclusion of the paper is that it substantiates a strong genetic affinity between the Red Deer Cave people of Southwest China ca. 14,000 years ago, and the main founding population of the Americas which broke away from Asia at roughly the same time or a few thousand years earlier.


Her mtDNA was a now extinct basal branch of mtDNA M9, a non-African modern human mtDNA clade, that is fairly closely related to the mtDNA M9 clades that are common in China's interior and central Asia (and to a lesser extent in the rest of China). Her mtDNA deviates by just one mutation from the mtDNA M9 root.

In island Southeast Asia and Vietnam, in contrast, mtDNA E which is also a basal branch of mtDNA M is more common.

Autosomal DNA

She is definitively an anatomically modern human based upon her autosomal DNA, which is close on a PCA chart of overall genetic similarity to Hoabinhian hunter-gatherers who occupied Southeast Asia and adjacent regions in China from sometime in the Upper Paleolithic, a.k.a. Late Pleistocene, era until about 2000 BCE.  

There has been considerable genetic continuity from her to modern populations that were until recently hunter-gathers in the region over the last 14,000 years.

She shows an even greater autosomal genetic affinity to contemporaneous Northeast Asians on the brink of the Bering land bridge and to the oldest Native American ancient genomes, than she does with mainland Southeast Asian Hoabinhians.

Her autosomal DNA reflected most strongly an ancestral DNA component now associated with Southeast Asia with a significant minority of ancestry that was distinctively Northeast Asian, in addition to trace ancestry components that were distinctive African, specifically Native American, European or Papuan (and she has no distinctive South Asian ancestry component).

This indicates that 14,000 years ago, there was already differentiation and population structure between Northeast Asia and mainland Southeast Asia that had formed, and then produced admixed populations again, of which she was a member.

Archaic Hominin Ancestry

Her percentages of Neanderthal (1.27%) and Denisovan (1.29%) autosomal DNA admixture is similar to that other modern humans of that era and does not show enhanced or recent archaic admixture. Her Neanderthal and Denisovan ancestry percentages are broadly similar to modern Asian people from this region today.

This is what the paradigm would expect since there are no really definitive examples of archaic hominins in mainland Southeast Asia or China more recent than about 100,000 years ago. 

Evidence of a Neanderthal presence has never been found east of South Asia in places south of the Altai. 

The History Of Two Asian Genetic Variants

She lacks a derived genetic variant associated with light skin in modern East Asians that an analysis of many ancient and modern DNA samples in the papers concludes arose only in the early Holocene era (i.e. in the last 10,000 years) with the oldest isolated example from 7,500 years ago and some examples in the las 5,000 years and predominance in the last hundred years.

The paper also analyzes the emergence of the derived EDAR variant which is the source of some of the distinctively "Asian" phenotype. This arises many thousands of years earlier than the light skin genetic variant, but apparently the ancient DNA sample from this woman was missing the part of her genome that would have contains the EDAR variant, so we don't know if she had it or not. The oldest example of the derived EDAR variant is from the Amur region in Northeast Asia from about 19,000 years ago (although mutation rate estimates suggest it originated 30,000 years ago), and it is fairly widespread in East Asia, Northeast Asia, and the Americas from 11,600 years ago onward, becoming predominant in Southeast Asia, East Asia, Northeast Asia and the Americas by 4,900 years ago.

Both of these derived genetic variants that are distinguishing features of modern Southeast and East Asians post-dated the Last Glacial Maximum which led to essentially total population replacement in Northern Asia (and preceded any significant modern human population that left lasting traces in the Americas).

The Paper's Analysis

The discussion section of the paper generally sums up these themes and adds some additional cross-disciplinary speculation, most of which is plausible, although the bold underlined conjecture really has no solid support.
Southern East Asia harbors rich archaeo-anthropological sites with rich morphological diversities, including the ∼100 kya Zhiren Cave and ∼120–80 kya Fuyan Cave in southern China (although a recent study suggested much younger dates for these two sites), the ∼190–50 kya H. floresiensis, and the 67 kya H. luzonensis in Southeast Asia.  
Indeed, based on the published mitochondrial genome data, the matrilineal lineage diversity of the Late Pleistocene hominins in eastern Eurasia is quite high, including the reported Ust’-Ishim (45.0 kya, novel N), Salkhit (34.0 kya, independent novel N), Tianyuan (40.0 kya, basal B), Yana-old (32.0 kya, U), MA-1 (24.3 kya, novel U), LLR (11.0 kya, M27d), and MZR (14.0 kya, basal M9) in this study, many of which were lost during the post Pleistocene (<11.7 kya) human evolutionary histories. 
In addition, the hominin fossils from these archaeological sites, such as MZR and LLR, all exhibit rich physical anthropological diversities, some of which were thought to overlap with the morphological characteristics of archaic hominins and triggered proposals of different scenarios of human evolution in this area.
In this study, we provide compelling evidence that the Late Pleistocene MZR from Malu Dong in Southwest China is an AMH. The nuclear genome data indicate that MZR represents an early diversified AMH lineage in East Asia. The mtDNA of MZR belongs to one of the root matrilineal lineages of AMHs in southern East Asia. Identified as a novel basal M9 lineage, MZR may represent one of the extinct pioneer hunter-gatherers ancestral to millet and rice farmers in China who emerged in the Yellow River and Yangtze River valley during the Early Neolithic period. Additionally, we observe obvious stratification and substructure of ancient human populations between southern China and mainland Southeast Asia, an indication of already diversified genetic backgrounds of the Late Pleistocene populations in southern East Asia. . . .
It should be noted that the MZR morphological data are indeed informative in reconstructing human morphological diversity during the Late Pleistocene in southern East Asia. However, due to the limited human remains as well as the limited number of morphological traits, it would be hard to confidently reveal the identity of the studied subject. To this end, genome sequence data are critical for unequivocal species identification, quantification of genetic introgression, and reconstruction of population history.
Spatiotemporal tracing of mutations related to phenotypic changes in human populations can help reconstruct the prehistoric patterns of how natural selection has shaped these adaptive events. We observed that the OCA2-HiS615Arg (rs1800414) variant, a key adaptive mutation causing skin lightening in East Asians, initially emerged in the southern coastal region of China during the early Neolithic (Liangdao2-7.5kya). The rapid dispersal of this variant during the past 4,000 years in East Asia is consistent with the proposed Darwinian positive selection on the adaptive allele (OCA2-615Arg), leading to skin lightening in East Asians to cope with the relatively low UV radiation in high-latitude areas. Interestingly, the rapid explosion of OCA2-615Arg coincides with the known major population expansion in China during the Late Holocene epoch. It should be noted that due to the limited aDNA data in East Asia, the inferred time of selection onset for OCA2-HiS615Arg is a rough estimation. With more aDNA data available in the future, we expect more accurate time estimation and high-resolution spatial-temporal tracking of adaptive genetic variants in East Asia.
Consistent with the dating of MZR (14.0 kya), following the end of the LGM (26.5–19.0 kya) and the earliest securely dated sites in Beringia (15–14 kya), we demonstrate that MZR has a higher affinity to First Americans than to Tianyuan (40 kya) and all the pre-LGM Late Pleistocene Siberians. MZR, Amur-19K, and UKY are cladal with respect to First Americans, while Amur-14kya and UKY exhibit a higher affinity to the Americans compared to MZR. Thus, MZR is linked deeply and indirectly to the ancestry that contributed to First Americans. We speculate that during the Late Pleistocene, there was an express northward expansion of AMHs starting in southern East Asia through the coastal line of China, possibly by way of the Japanese Islands, and eventually crossing the Bering Strait and reaching the Americas. However, the scenario that MZR shows a higher affinity to Americans compared to Jomon likely reflects that the 2.6 kya Jomon population does not represent the early post-LGM humans who settled in the Japanese archipelago. 
The proposed migratory route along the east coast of East Asia by way of the Japanese Islands is supported by a recent finding of a Paleolithic site (∼16 kya) at Cooper’s Ferry of western Idaho, USA, where they found the use of unfluted stemmed projectile point technologies before the appearance of the Clovis Paleoindian tradition. Notably, it exhibits early cultural connections with Paleolithic nonfluted projectile point traditions in Japan. The bifacial point and backed blade technologies (∼22–16 kya) in Honshu, Japan, lend technological correlates to the shared ecological and geographical factors with the Americas.
This scenario is also supported by the current distribution pattern of the ancient Y-chromosome lineage Hg C in coastal East Asia, Siberia, and North America. Approximately 40 kya, stemming from southern East Asia, the Hg C carriers started a northward expansion along the coastal regions of mainland China, the Korean Peninsula, and the Japanese archipelago, reaching Siberia ∼15 kya, and finally made their way to the Americas.  
In addition, unlike all other East Asian populations, the indigenous Ainu people in northern Japan and Sakhalin Island, Russia, show a closer genetic affinity with northeastern Siberians than with central Siberians. Hence, the Japanese Islands may serve as the midway station along the proposed migratory route, and aDNA data of Late Pleistocene human remains from Japan will be highly informative in testing the proposed coastal route. 
Finally, the spatial-temporal distribution of the East Asian-specific EDAR-V370A variant, as well as its early presence in the LosRieles-12.0kya sample from coastal Chile of South America, supported a clear contribution of the Late Pleistocene East Asians to the first Americans.
In summary, we generated ancient genome sequences of MZR, a Late Pleistocene female who lived in Southwest China, one of the global biodiversity hotspots and the ice age refuge region. The aDNA data confirm that she possesses diverse genetic components and represents an early diversified population, suggesting the scenario of more diverse AMH lineages than previously thought during the Late Pleistocene in southern East Asia. Our study paves the way to explore genetic explanations of morphological complexities of early hominins. MZR also shows a deep and indirect link to the ancestry that contributed to First Americans, which may help reconstruct the earliest migratory route from East Asia to the Americas.


Ryan said...

1% Denisovan is actually quite a bit higher than normal.

Tom Bridgeland said...

Sadly, the COVID pandemic and the unreliability of research results from China has soured me on all Chinese research results. Nearly everything they told us to do was disproven.

andrew said...


The linked popular treatment of an older article says it's !% Denisovan is normal for mainland Asia (I thought it was high and checked for a source), and the paper does too with citations (around footnote 37) and I clicked through to the source which wasn't as clear as I'd like so I looked for another source. Partially, it is an issue with your methodology for determining the percentage, I think, as well. It also says that the total archaic introgression is in the normal range.


My experience with Chinese research papers in a variety of fields including genetics, is that they tend to be monodisciplinary to a fault, ignoring obvious issues from related fields, and has its share of ESL English ineloquence, but isn't otherwise concerning in integrity.

For example, this paper is actually contrary to the multi-regional hypothesis which was the Chinese Communist Party preferred position for a long time (although it certainly doesn't call attention to that fact).

neo said...

sounds like contamination with modern dna

Tom Bridgeland said...

neo, I immediately thought the same.

andrew said...

Contamination with modern DNA would lower the Denisovan percentage or leave it unchanged, not increase it, in most scenarios, given the demographics of typical aDNA researchers.

Otanes said...

Bummer they weren't super archaic-admixed. There's something entertaining about the idea of such a narrative.

andrew said...


If wishes were horses . . .