Australia: The Land Where Time Began

A biography of the Australian continent 

Interbreeding with Denisovans in Oceania

There is an alternative scenario that Green et al. (2010) didn’t consider for that publication, though as co-authors, they subsequently offered in the context of the hominin phalanx and molar tooth, Middle Palaeolithic/Upper Palaeolithic transitional context in Denisova Cave, Southern Siberia (Kraus et al., 2010; Reich et al., 2010) that were recently sequenced, is the possibility that AMH interbred with other non-Neanderthal populations, such as H. Heidelbergensis, or descendants of H. erectus that were moving from China to Central Asia.

In the first report of the Leipzig group (Krause et al., 2010) argued from the complete mtDNA of the phalanx that the Denisovan lineage had long ago branched off (1.04 Ma, CI: 0.779-1.3 Ma, so that the Neanderthal and the modern human linages are more closely related to each other (0.493 Ma, CI:  0.3744-0.6121 Ma) than either were to Denisova.

They reported in their second publication another complete mtDNA sequence, that was from the tooth, which proved to be very close, though not identical, to that of the phalanx that had been recovered from the same site, and this indicated the presence in the cave of 2 ‘Denisovan’ individuals that were very similar to each other at the mitochondrial locus, and which confirmed the phylogenetic age estimates, with a slightly reduced error (0.982 Ma, CI: 0.7805-1.208 Ma. The Denisova tooth was shown to group tightly with early Homo specimens morphologically (Australopithecines, H. habilis, and African H. erectus) all of which had teeth that were significantly larger than a group that was more modern that contained Neanderthals, H. heidelbergensis and all other European humans, both ancient and modern, all AMH, and H. erectus from China. Specimens from H. erectus in Indonesia tended to be intermediate between the 2 groups. A massive, splayed lingual root is another non-metric feature that grouped the Denisovan tooth with the erectus types. For 2 other genetic comparisons, in offline material, the Leipzig group displayed Venn diagrams of autosomal segmental duplications that show that those held in common, first between a human of the present (NA18507), a Neanderthal, Denisova and a chimpanzee, then between Neanderthal, Denisova and a human from the present. The modern and Neanderthal in each case shared most segmental duplications with each other and least with Denisova (Reich et al., 2010).

According to Oppenheimer these 4 comparisons, 3 of which were genetic and 2 morphological are all consistent with each other, first in making the Denisova the biological outgroup when it was compared with the European group that were related closely to each other (AMH, Neanderthals and H. Heidelbergensis), though further it could imply that Denisova could have descended from an earlier African population. Martinón-Torres (2010) has argued the case for a Eurasian origin for Denisova.

In the second Denisovan paper (Reich et al., 2010) also reported extensive sequencing which they claimed shows that 4-6 % of provisional autosomal signatures obtained from the Denisovan finger bone were found in 3 Melanesians, 2 Papuans and a Bougainvillean, though not in any of the other groups of modern humans they tested, African: San, Yoruba and Mbuti; 5 “Eurasians”: French, Han Chinese, Sardinian, Cambodian, and Mongolian; and South American: Karitiana. The inclusive or exclusive geographical scenario could be consistent with ancestral Oceanic AMH mixing, not with H. erectus from North China, but with H. erectus in the region of Indonesia of the present, given all the genetic and morphological observations summarised so far.

It has been suggested that ‘Solo Man’ was a version of H. erectus that was significantly larger brained than H. erectus from Ngandong and Sambungmacan in central Java, may have survived to only 27 ± 2 to 53.3 ± 4 ka (Swisher et al., 1996), and as this was controversial the regional distinction is important. Some of the same researchers revisited the site because of the high degree of controversy about the dates obtained the first time, the results of this study were dates that were higher, but were conflicting, the youngest being an ESR/U-series date of 143 ka (+20/-17) (Indriati et al., 2011; see Dennell Ch. 4). Oppenheimer suggests that whether or not the first dates were correct these evolved H. erectus descendants could be brought to the Upper Pleistocene.

It was argued for a closer descent relationship between Neanderthals and the Denisovan hominin, than either had with AMH (Reich et al., 2010), though in view of the erectine links and the range of the genetic and morphological evidence, which Oppenheimer found surprising. The change in emphasis was based on another comparison, autosomal SNPs (single nucleotide polymorphisms) between Denisovans, chimpanzees, and Neanderthals, though based on the relative degree of sharing of SNPs that were distinctive to Neanderthals. According to Oppenheimer the results and implications of the latter were different when compared with the previously mentioned 4 genetic and morphological comparisons, and the simple explanation for this is that they didn’t test for Neanderthal-Denisovan admixture.

He suggests the strength of the evidence their assertion is based on for more recent common descent for the 2 archaic groups was not as convincing as that for the simple intrusion of Neanderthal into AMH. According to Oppenheimer this is the result of not testing a more likely explanation for the autosomal SNP associations between Neanderthals and Denisovans in this more complicated 3-group interbreeding puzzle. Prior to the arrival of AMH, that would have been extended direct hybrid interbreeding between the 2 archaic human groups locally in Central Asia, as supported by their geographical, physical and temporal overlap, which is consistent with the presence of Neanderthal mtDNA in that part of Central Asia (as shown in Okladnikov Cave, not far from Denisova Cave; Krause et al., 2007b). Oppenheimer suggests interbreeding would constitute only admixture, of course not recent descent, and could be the explanation for the extraordinary plesiomorphy as seen in the “Mongolanthropus” that was discovered recently at Salkit in Mongolia. This skullcap fragment, that has marked superciliary arches “shows multiple similarities with Neanderthals, Chinese Homo erectus, and west/Far East archaic Homo sapiens” (Coppens et al., 2008; but see also Kaifu & Fujita, 2012), who mention its dating of 20,000 BP and regard it as within the range of Late Pleistocene AMH.

The comparison of autosomal SNPs was in Denisovans with a variety of those from a number of different Eurasian populations in the paper by Reich et al., (2010) was the result that Oppenheimer found to be most interesting which revealed that Denisova shared 4-6 % of its genetic material with Melanesians of the present in the Pacific, which are a quarter of the globe away from Denisova Cave, though not with any other population that was closer to the Cave. Other research (Rasmussen, 2011) has shown that there is the possibility of archaic admixture in Oceania was reinforced by the discovery of Denisovan admixture in an Aboriginal man [from the far southwest of Western Australia who allowed a lock of his hair to be taken 100 years ago, before European colonisation reached his area] that was similar in degree to that found in Melanesians. It was also postulated (Rasmussen et al., 2011) that 2 dispersals of AMH populations that were African-derived, into Australia: the first occurring 72,000-62,000 years ago, and a later dispersal 38,000-25,000 years ago. Oppenheimer suggests their interpretations were influenced by selective use of autosomal-dating of demographic events, which had not been supported archaeologically. Oppenheimer suggests that their analysis and interpretation of the phylogeny of ancestral populations depends, as in the case of Green et al., (2010), on a statistical association test that is based on only 4 complete genomes, and not on any genetic phylogeny.

Oppenheimer suggests the results of (Rasmussen et al., 2011) still extend the ‘Denisovan DNA’ influence in Oceania more broadly into the Sahul region (as predicted, Oppenheimer, 2012b) and do not falsify the consensus of a single AMH exit model that is the preferred option in this book¹. As Oppenheimer points out there is skeletal evidence that would reverse the morphological order of their “two Australian waves” as the earliest human crania known from Australia are gracile and more modern in appearance, though robust skulls at the Kow Swamp and Coobool Creek sites have an archaic appearance, and have a radiocarbon date of at most 14,300 years ago (Brown, 1992).

A more recent study (Reich et al., 2011) has extended this antipodean problem by reporting an extra 33 more locations in Asia, Southeast Asia and Oceania. The 15 populations tested on the eastern side of Huxley’s line, which include 1 Negrito and 1 non-negrito group, in the southern Philippines, Near Oceania and Polynesia, show significant evidence of Denisovan intrusion, as high as, but no higher, than that found in Melanesians (i.e. near Oceanians), and now a similar level has been found in Australians. To the west of the Huxley Line 27 Asian and Southeast Asian (SEA) populations, that included 2 Negrito groups, none show any significant evidence of such intrusion, according to Reich’s analysis. Therefore it appears all admixtures occurred offshore in Wallacea or Sahul, as the true eastern limit of the Asian mainland up to 10,000 years ago. It is possible more westerly admixture with later extinction is possible, though it is much less likely.

Oppenheimer suggests relevant questions about these findings are how many times, from where, and when? In the case of the Philippines the results of relative admixture are all consistent with a single proximate common source of Denisovan admixture: Wallacea, Australia or New Guinea along with subsequent fresh external diluting with AMH gene flow. In Wallacea, Oceania and the Philippines the non-Denisovan admixture analysis shows that New Guinea and Australia are tightly correlated, major alternative candidate sources of  AMH gene flow for the Pacific Region, though Australia is consequently the richer one, overall by 40 %. If Near Oceania was the primary dispersal source, the apparent anomaly could still be compatible with it, though with the New Guinea highlands (Gosden, 2010; See Summerhayes & Ford) ¹ , having gone through founding event(s), isolation and subsequently drift, a scenario that is inferred from shared human leucocyte antigen (HLA) markers (Serjeantson & Hill, 1989). Based on geographic considerations Wallacea (Nusa Tenggarah and/or Moluccas) still seems the most parsimonious Denisovan source. A plot of individuals’ Denisovan admixture against Near Oceania (New Guinea) admixture has been constructed (Reich et al., 2011) which shows tight correlation and high values of Denisovan admixture (ranges from 30-100 % for Wallacea and Fiji, and this speculation is consistent with that plot. Polynesians, though they group together with the lower end of the distribution for Wallacea, with lower values of 20-30 %, and still correlating with admixtures as found in Near Oceania, have been shown to have consistently less Denisovan intrusion than they would be expected to have had from that admixture, which Oppenheimer suggests is a likely result of later Southeastern Asia admixture and drift.

When compared with the other 13 sites the south Philippines are anomalous, as 3-4 times more Denisovan intrusion is shown than would be expected from their Near Oceanian (i.e. non-Denisovan) admixture values and no clear correlation with them. A separate admixture event and subsequent dispersal is probably indicated by this, though north Wallacia could still be the source region.

It is suggested in Melanesians and Australians, the similar, asymptotic Denisovan intrusion that the main admixture event occurred in a single source population before Sahul was colonised, and this probably occurred in Wallacea. The paucity of regional samples from Sahul, which could allow for subsequent migrations or even admixture events, is a caveat on the above fresh inferences.

Sources & Further reading

1. Stephen Oppenheimer in Dennell, Robin & Porr, Martin, eds., 2014, Southeast Asia, Australia, and the Search for Human Origins, Cambridge University Press.