Category Archives: hybridization

Anomalous Paleolithic skull from South China

This is sure going to cause some heated debates:
See also the press release (Eureka Alert) and a photo gallery (Live Science).
And this is what is all about:

It’s not the skull of Darth Vader, although I know you thought so as soon as you saw it, but actually that of a person living some 14,000 years ago at Longling cave, known as Longlin 1.
The authors of the paper argue on craniometric grounds that the specimen may not be a modern human, however in all PC analysis the skull falls (oddly enough) within or very close to modern human clusters. Only in the PC3 some differences show up. 
In fact the closest among those skulls analyzed in fig. 8 is Mai Da Nuoc[PDF], an Early Hoabinhian skull from nearby Vietnam, which is claimed to be Australoid (but does not look at all like the Australian skulls I’m familiar with, normally quite broader).
In fact I have been searching and I found at least one skull that looks a lot like Longlin 1: Peñón woman from Mexico:

She’s one of a number of Paleoindian skulls that have puzzled prehistorians because they are not quite like most modern Native American skulls. However to my eyes, the Peñuelas skull (Chile) skull is not that different, even if it is also more like modern Native American ones.
Also to my eyes, those very marked cheek bones (regardless of whether they may have been produced by mastication) remind me of one of the most characteristic traits of the so-called Mongoloid type (which is quite unreal and plural in fact): prominent cheekbones that give the face a flat appearance.
But I’ll leave that to the experts. Not worth bumping heads for something that may well be partly epigenetic/environmental for all I know.
What I do not see is any ground for the claims of the authors that it could be a newly found species. They even have worked a fancy reconstruction that makes Longlin 1 to look like a well-known reconstruction (by the same author) of the notorious Hobbit (when the skulls do not look at all similar):

Credit: Peter Schouten (info)
Notice please how any trait that could make him look Mongoloid (East Asian or Native American) has been scrapped from the reconstruction: he has dark brown skin, beard (which could well have been shaven by a professional just yesterday) and a hairy body, a very broad nose that does not fit at all with the small triangular orifice, absolute lack of epicanthic fold – plus very small ears that make him look odd. 
Another analyzed skull from the same area is Maludong 1704, of which only the vault remains. This one overlaps in the analysis (fig. 9) with Crô-Magnon 1, however when more samples are added (fig. 10) some archaic skulls (H. erectus) also show up close, as do again CM1, early African H. sapiens and Nazlet Khater 2. 
For me, with due caution, modern proto-Mongoloid H. sapiens (anomalous) but your call: the debate is served.

Update (Mar 17): I must mention that the excellent anthropological artist Zaender (which I have mentioned before) has also produced his own versions of the Maludong people without most of the exotic fancy of Schouten’s reconstruction:

Credit Zaender

He has another example at his blog: Regional Ancestry Bands.

The ears still look small to me and the nose and lips unnecessarily too wide but it is probably more correct re. eyes and hair.

In any case it makes evident that it’s very difficult to actually reconstruct a face from just a skull when all the expressiveness and even most ethnic traits are in the flesh and skin, being impossible to discern from a mere skull.


Splitting hairs with the Neanderthal affinity

John Hawks published today an interesting albeit potentially misleading exercise of comparing (known) Neanderthal DNA (Vi33.16)  to moder humans by HGDP samples.
The first graphic is a for example a very visual representation of why geneticists have concluded that there is a percentage of Neanderthal admixture in non-African humans, a striking visual synthesis of the results of Green 2010 with other modern samples:

It is easy to see for all in this graph that if the median shared Neanderthal variants in Africans is c. 626,000, while in Eurasians is c. 644,000 (visual estimates), then there is something going on and admixture is the most likely explanation.
A simple cross-multiplication exercise shows that the admixture apportion using these medians would be c. 2.9%. However a cautionary use of a higher figure among the African variability range (likely not caused by admixture but retained ancestral diversity) such as 630,000 yields 2.2%.
Green 2010 and later reanalysis by the NGP team estimated 2.4% (although they initially talked of 1-4%), all of which illustrates how is not easy to come with an exact percentage figure and that some uncertainty remains and must remain by the very nature of the exercise and the samples involved.
Splitting hairs
But of course we love to split hairs, at least a bit. I must admit I did it myself back in the day with the handful of samples used by Green et al. originally. Then I was asking rhetorically: are Chinese slightly “more Neanderthal” than other Eurasians?
Not quite because of the uncertainty implied in all the comparison is the real answer: the apparent differences are too small to be significant.
And this more or less what Hawks ponders in his article. As you can see above Europeans appear now slightly more Neanderthal than East Asians. However the difference is actually trivial: approx. 1000 base pairs, what is a variance of 0.12 percentile points of that approx. 2.4% (Hawks writes ‘half-percent’ when talking of intra-European differences of the same range, but he must be measuring something else than I am: 0.005×2.4%=0.012%, maybe he meant 5%… of that 2.4%? Unsure and, as he does not allow comments, I can’t ask).
However, in spite of formally acknowledging this insignificance of the differences, he goes on to state the following unlikely hypothesis:

At present, we can take as a hypothesis that Europeans have more Neandertal ancestry than Asians. If this is true, we can further guess that Europeans may have mixed with Neandertals as they moved into Europe, constituting a second process of population mixture beyond that shared by European and Asian ancestors.

While it’s not absolutely impossible, the data does not support any meaningful extra admixture in Europeans but actually what it does support is the lack of any significative difference through Eurasia. IF there was any extra admixture in Europe (or better West Eurasia, what’s the obsession with Europe?) it is not detectable and hence was surely hyper-minimal.
We are therefore before yet another case of wishful thinking, of which I have stumbled upon several, much more severe cases in the las weeks alone. The illusion of a Neanderthal admixture or assimilation or even full continuity into, specifically, modern Europeans (usually West Asia is totally ignored even if it was there where most of the Sapiens-Neanderthal interaction must have taken place) is an obsession difficult to put aside for some I am learning.
Hawks is still quite serious and scientific and knows the ropes of genetics quite a bit and, therefore, he does not insist on that too much, showing different angles and comparisons that are interesting albeit unsupportive of his outlined hypothesis. However he does not abandon that unlikely boat so obviously sinking.
I am realizing that it is much harder for some Eurocentric multiregionalists to abandon their old Neanderthalist hypothesis than I would have expected. After all the genetic data is there for all to see and I must say that Hawks provides us with highly informative eye-candy here, which clearly supports the Neanderthal admixture episode and the uniformity of it across the various Eurasian populations.
Yet he seems blinded by the C.I. variation, which is caused, no doubt, mostly or only by local founder effects and/or drift (after the Neanderthal admixture episode).
Illusion of African Neanderthal admixture
A good example is again provided by Hawks himself:

Yorubas here appear some 2000-3000 BPs more Neanderthal than Luhyas. Even Hawks admits to be puzzled by this result, which is obviously attributable to mere ancestral diversity within Homo sapiens (but not for him). He expected the opposite result: that Luhyas, who live in Kenya, would be more influenced by Eurasian back-flow into Africa and display some greater Neanderthal affinity. 
Actually this comparison does not just illustrate well how such small ranges of variation are normal and a remnant of ancestral diversity within the species (most probably) but also illustrates how Dienekes’ hypothesis about Neanderthal admixture being in fact internal structure of Homo Sapiens before the migration out of Africa (or even after, because he has also argued for a greater role for Arabia and what not) is a total fantasy. 
Thanks for the interesting and beautiful graphs, Dr. Hawks, but I cannot agree with your hypothesis because I see zero support for it in your own data. I think you have a clear case of splitting hairs syndrome, probably a symptom of repressed multiregionalist grudge (Eurocentric Neanderthalist variant).

‘Denisovan’ admixture may actually be from H. erectus

H. erectus (fem.) reconstruction

Neanderfollia mentions today[cat] a new paper published at the moment only at arXiv that re-analyzes the data from Reich 2010 and related papers and concludes that the Denisovan admixture may well be original from Homo erectus, and not even individuals closely related to the Denisova cave specimens after all.

This article shows how to fit reticulate finite and infinite sites sequence spectra to aligned data from five modern human genomes (San, Yoruba, French, Han and Papuan) plus two archaic humans (Denisovan and Neanderthal), to better infer demographic parameters. These include interbreeding between distinct lineages. Major improvements in the fit of the sequence spectrum are made with successively more complicated models. Findings include some evidence of a male biased gene flow from the Denisova lineage to Papuan ancestors and possibly even more archaic gene flow. It is unclear if there is evidence for more than one Neanderthal interbreeding, as the evidence suggesting this largely disappears when a finite sites model is fitted.

The paper needs some style revision but is otherwise very interesting, even if it’s largely an exercise of statistical analysis, often resulting somewhat arid.

Some excerpts:

(…) one of the most surprising features of the planner NeighborNet model is that it does not reverse the positions of Neanderthal and Denisova, so that Papuan could have a unique split with the Denisovan (as Reich et al. 2010 suggest the Papuan lineage received ~5% of its genes from that lineage). As we will see later, the apparent reason for this would seem to be that the distance from Denisova to Chimp is more strongly underestimated than that from Denisova to Papuan. The underestimation of the Denisova to Chimp distance could be due to Denisova harboring some very archaic alleles, or it could be sequencing error.


The decrease in frequency of the DP pattern on X, particularly when compared to the NP pattern (which is near autosomal average frequency on X) suggests the possibility of asymmetric gene flow in this introgression event. If so, it would seem that this might be most readily explained by greater survival and reproduction of the offspring of Denisova males impregnating the modern human female ancestors of Papuans rather than the other way around.


Note the high frequency of the DNP pattern, which may be due to the Denisovan relatives that mixed not being closely related to the Denisovan sampled.


It seems tempting then to think that a model of three independent out of Africa lineages, with three independent mixings with the same population of Neanderthals (plus the independent Denisovan mixing event), would fit markedly better than the present model.

This last bit I find hard to believe, notably because we know of no Neanderthals ever existing in East Asia. In addition it would be a very odd coincidence that all three arrived to the approximate same amount of Neandrthal admixture. It seems much more likely that these smaller differences have been fixated as the three populations diverged in the Greater Eurasian expansion, after the OoA initial migration, when they probably incorporated the Neanderthal ancestry. (My two cents anyhow).

In any case, their conclusions follow:


Overall, the fitting shows that a hierarchical structured coalescent model with at least two introgression events between archaic humans and out of Africa Moderns leads to a substantial increase in fit. Overall fit however, is still far far worse than could be expected. It seems that to improve the fit a number of factors may come into play. Firstly, there are too many private NH, NF and NP [Neanderthal-Han, -French and -Papuan] patterns. Secondly, the latter of these, NP, seems markedly less than the former two. Thirdly, there may be too many sequencing/alignment errors in the present data to confidently move towards refining so many parameters and the overall fit. The marked improvement in fit when a finite sites model is employed is consistent with this. One model that may do a better job of describing the data with fewer parameters is independent mixing of Neanderthal genes with Han and French, but to a nearly identical total degree. Also, lesser mixing of Neanderthal genes into Papuan, made up for by a larger proportion of archaic alleles in Papuans coming from the mixing with an archaic that is only slightly closer to Denisova than to Neanderthal. This would in turn suggest that the mixing with Neanderthals was not purely right out of Africa and it was not a single event. Instead, there may have been opportunity for European ancestors to pick up Neanderthal alleles, in the unknown part of Eurasia they existed in prior to moving into Europe, ditto and independently for the ancestors of the East Asians, while Papuan ancestors moved fairly rapidly through the zone of classical Neanderthals and picked up most of their archaic genes in the Indonesian region. The form of this ancestral population may have been about equally related to Neanderthals and Denisovans, but may also have had an appreciable proportion of even earlier (e.g., Homo erectus genes) in its genome. This last point comes up in a number of analyses including the resampled NeighborNet and the finite sites model, but confirmation is difficult as the rate of sequencing / assembly error could be having a similar effect.

For background in this blog and its antecessor Leherensuge (from oldest to most recent):

Minimal ‘Denisovan’ admixture in SE Asians

In outright contradiction with a recent paper by Reich, Skoglund and Jakobson now propose that there is tiny bits (c. 0.67%)  of Denisovan admixture among SE Asians, notably South China populations.
Pontus Skoglund & Mattias Jakobsson, Archaic human ancestry in East Asia. PNAS 2011. Open access. [doi: 10.1073/pnas.1108181108]
I must say that I really hate these highly complex papers that use ill-explained statistical modeling: they make almost impossible to make a criticism of any sort and make me suspect, right or wrong, that they could be burying, consciously or not, inconsistencies under the complex modeling.
But they could be right and I’m just to shallow to understand.
In any case, it could make some sense with a pattern I did detect in Reich 2010:

In (…)  table S8.2 French, Han and Cambodians (and only them) also appear to show some admixture with Denisovans, though maybe a third or fourth of that of Melanesians.

I almost nailed it according to this paper. 
It would also be consistent with the alleged ‘Denisovan’ introgression of HLA*A1. But this introgression was called to question, as the allele is common in Uganda however, what is inconsistent with any introgression in Eurasia and rather suggest random founder effects (or selection) from an African source.
So your take, I have no opinion – at least not a clear one. 
Thanks to Neanderfollia[cat] for calling my attention on the matter.

Update: take a look at John Hawks’ opinion if you wish. It’s interesting.


‘Denisovan’ admixture widespread beyond Wallace Line, non-existant elsewhere

Reconstructed H. erectus
Remember that last Christmas we got an unusual gift of knowledge in the finding by Reich et al. that Melanesians of all modern humans researched back in the day were the only ones to show admixture with the mysterious Denisova fingers?
Remember that I said already back then that this admixture was not with Denisovans as such but a related species (probably H. erectus) of which the Denisova hominings were just the tip of the iceberg and a Neanderthal-admixed tip actually.
I proposed therefore that the admixture shown by Melanesians but not continental Eurasians was probably the product of admixture with H. erectus solensis (or something like that) in Indonesia, while ‘Denisovans’ were hybrids of H. erectus and H. neanderthalensis, possibly at near 50% levels.
I suggested then this scenario:

With the 2nd admixture representing this regional H. erectus introgression and the 1st one being that from Neanderthals or maybe a related Heidelbergensis-derived population in South Asia (Hathnora hominin).
I also said that the figures of ‘Denisovan’ admixture had to be cut by half because the authors were counting Neanderthal admixture twice in Melanesians (as ‘Denisovans’ were probably Neanderthal-Erectus hybrids). Quoting myself:

They suggest (supp. info 8) that Melanesians would have as much as 7.4% of admixture with archaic species: 4.8% Denisovan plus 2.5% Neanderthal. But, if Denisovans are hybrids of H. erectus and H. neanderthalensis (as seems most likely, see above), then the real admixture with H. erectus would be an undetermined percentage but always less than 4.8%. As we know that the Neanderthal (or Heidelbergensis) component is 2.5%, it is most likely that the actual Erectus admixture in Melanesians is of only 2.3% or 2.4%, totaling 4.8%.

Now we are told that all the aboriginal peoples of Near Oceania, plus Wallacea and Filipino Negritos, show that admixture at similar or lower levels:
I could browse the paper a day or so ago, so I hoped this was an open access paper. Yet today I find it is PPV. Luckily Dienekes has published most of the relevant graphs at his blog.
In any case the relevant information is this map (from Neanderfollia[cat]):

It tells us that Papuans and Australian Aborigines share the greatest fraction of ‘Denisovan’ introgression, followed by Boungaville Melanesians, Fijians, Timorese, Alorese and Mamanwa speakers (probable Ati). These and other peoples of beyond what used to be the continental landmass of Asia in the Ice Age, retain some level of ‘Denisovan’ admixture. 
But Denisova is very far away and no admixture is known elsewhere. Why? Because the admixture surely happened in or near Indonesia and was not with the Neanderhal-hybridized Denisovans but with pure H. erectus from the region.
Papuans and Australian aborigines have probably 2.4% admixture from H. erectus, in people like the Timorese that would be 1.2% and in a group like the Roti (RO) it is of just 0.6%. That’s my interpretation of the available data. 

But from Sundaland to the West and North there is no such admixture: zero!
And that can only be explained if the admixture happened in Indonesia, maybe in Flores?

Strong introgression of non-Sapiens antigens in ‘Eurasian’ modern humans (but not in Africa)

HLA-A11: the legacy of H. erectus in ourselves
Please read the updates below because some of the claims may not be sustainable (however others are plausible).

Prof. Peter Parham (Stanford University, USA) has found that some of the most common immunologic alleles among non-African modern humans have been adopted from other species of Homo living in Eurasia upon the migration out of Africa. 

This would be a typical case of introgression, a process in which, by means of lesser admixture highly adaptive alleles from another population are adopted. The logic here is that Neanderthals and other Eurasian hominins would be much better adapted to Eurasian-specific diseases, while our species would initially be  adapted to African-specific diseases instead. 
According to New Scientist:
One allele, HLA-C*0702, is common in modern Europeans and Asians but never seen in Africans; Parham found it in the Neanderthal genome, suggesting it made its way into H. sapiens of non-African descent through interbreeding. HLA-A*11 had a similar story: it is mostly found in Asians and never in Africans, and Parham found it in the Denisovan genome, again suggesting its source was interbreeding outside of Africa.


Half of European HLA-A alleles come from other hominins, says Parham, and that figure rises to 72 per cent for people in China, and over 90 per cent for those in Papua New Guinea.

The dominance of Denisovan alleles in Eastern Eurasia is coincident with my theory of these Denisovan specimens being hybrids of H. erectus and Neanderthals and acting actually as a proxy for H. erectus, with whom some of our ancestors would have hybridized in SE Asia, where this hominin is known to have existed until very late dates compatible with our arrival to the region.
Besides HLA, only Melanesians show some clear (albeit very minor) Denisovan admixture, but in the realm of antigens, the legacy of H. erectus has been clearly stronger. 


News found thanks to Neanderfollia[cat].

See also in this blog:

Update: John Hawks, who has also been studying HLA, has objections to the conclusion that these alleles come from Denisovans or Neanderthals. For what I could gather he has two objections:
  • Age estimates, which are high risk slippery terrain.
  • Insufficient resolution of the genetics of the archaic hominin genomes. 

Important Update (Jun 18): the “Neanderthal” allele probably Sapiens, the “Denisovan” one may stand:

Hawks (same post, updated or did I miss it in first read?) directs us to a database of allele frequencies through the World, which would seem a most useful reference site. There we get clear evidence that the “Neanderthal” allele HLA*C:0702 probably migrated with our ancestors from Africa and needs no introgression explanation at all. The allele is frequent enough in many African populations peaking among the Baka Pygmies with 15%.

More complicated is the case of the allegedly Denisovan (Erectus) alleles HLA*A11. A look at the database is very clear: no native African population (south of the Sahara) has it at all except the creole ones of Cape Verde and Sao Tome (where it has without doubt recent European origin) and, crucially, a sample from Kampala, Uganda, where it reaches 4.3%.

This sample one is the only one that could suggest an African origin for this set of haplotypes. It could be argued however that as some genetic back-flow from Asia exists in the area, this case is explained by genetic back-flow. However the apportion is rather high, almost as high as Moroccan Berbers, Italians or Macedonians. Even the largest possible Asian source (Omanis: 11.4%) does not seem to be large enough to justify this island of HLA*A11 in Kampala.

Additionally the Nilotic Nandi of nearby Kenya are reported to have 0% of the controversial alleles. It is really hard to explain how this island of HLA*A11 arose in Kampala. But on the other hand, the fact that it is such an isolated finding is equally suspicious: if the allele (essentially HLA*A11:01) was so old in Africa, we should expect it to be found at least a very low frequencies in other populations.

Fine that malaria or other tropical diseases may have played a contrary selective role, as Hawks argues, but still the allele could, should, have survived in populations not affected by this disease. Also Uganda is not less Malaria-prone than Kenya (or most other nearby countries). So this explanation is not satisfactory.

A very localized founder effect of Asian origin (or a reporting error maybe) would seem to be at the origin of this anomaly. Also no African presence of variants 02, 03 or 04 has ever been reported.

Of course, we must always await for further data and research, but on the grounds of what we have now, I would say that:

  • Claim 1: HLA*C0702 is a Neanderthal introgressed allele. Busted!
  • Claim 2: HLA*A11 (several alleles) is a Denisovan (Erectus probably) introgressed allele. Plausible (but watch that Kampala island in Africa). 

Update (Aug 26):

The reference paper is:

Laurent Abi-Rached et al., The Shaping of Modern Human Immune Systems by Multiregional Admixture with Archaic Humans. Science, 2011. Pay per view

The supplementary material however is freely available


Explaining ‘Denisovan’ and also ‘Neanderthal’ admixture: the simplest scenario

I have recently discussed the Denisovan admixture in Melanesians discovered by the Neanderthal Genome Project and I discussed back in its day the Neanderthal admixture in all non-Africans (see here and here).
While the Neanderthal admixture episode may be easy to explain and was thus explained by Green et al. as happening early in the migration out of Africa, probably before arrival to South Asia. The Denisovan admixture in islands far away from Altai is not so easy to understand and has not been satisfactorily explained by anybody I know so far.

What were the Denisovans?
Denisova cave
First of all we have not a very clear idea of what kind of hominin were the Denisovans. Well, actually we know that their tooth clusters with Indonesian H. erectus, H. habilis and australopithecines (but also with the H. sapiens of Pestera cu Oase, quite divergent from the rest in this aspect).
We also know that the Denisovan mtDNA belongs to a branch older than that of H. ergaster and descendants, because it is almost twice older in its divergence from that of Neanderthal and Sapiens mtDNA (both derived from H. ergaster c. one million years ago by all accounts that make any sense). What diverges in the common tree of Humankind (senso lato) almost twice that time? Asian H. erectus, believed to derive from a population represented by H. georgicus.
Nothing else does. Hence the Denisovan mtDNA, found in two different individuals (a finger and a tooth actually) must be that of Asian H. erectus.
However the Denisovan nuclear DNA is not so distant from Neanderthals. What does it mean? Most probably that they were a hybrid Erectus-Neanderthal population, what fits well with their presence in Altai (at the crossroads of known homelands of both species), their use of Mousterian technology (typical of Neanderthals) and the presence of Neanderthals in similar dates at nearby sites.
So my theory about Denisovan identity is this one: they were a hybrid population of Neanderthals and H. erectus, with maternal lineages of the latter species and technology of the former.
Melanesians in Siberia? No way!
blond Melanesians
Quite obviously Melanesian ancestors were never in Siberia. This is not just a matter of the coastal migration model, that also, but specially a matter of pigmentation. The name Melanesia means Islands of the Blacks in modified Greek and, if the ancestors of these peoples would have been in Siberia for any extended period, they would have lost their tropical pigmentation for sure because otherwise they would not be getting enough vitamin D and their children would be extremely unfit for that reason (retarded, schizophrenic, rickety, etc.) And, as the case of Native Americans clearly illustrates, re-evolving black pigmentation, once it is lost, is no easy matter. In maybe 15,000 years tropical native Americans have only got a tan.
So the ancestors of Melanesians and other very dark tropical Asians have definitively not lived in Siberia at any time. Besides, it is totally non-parsimonious in what regards to modern human mtDNA and Y-DNA spread, the tropical route is much more logical and natural.
So they must have admixed with some relative of Denisovans elsewhere, for example in Sundaland, where some Homo erectus are known to have lived in dates that are perfectly compatible with this scenario. An encounter of the first of our species arriving to that area and Homo erectus soloensis is almost sure to have happened.
So we do have a plausible and even likely scenario for this admixture event in the ancestors of Melanesians, not in Altai but in SE Asia.
Admixture detection by proxy… interesting.
Certainly that we can detect admixture happening in Java by studying distant relatives in Altai is interesting. And it makes sense. If you compare a modern French-Vietnamese with French and Altaians it’s likely that he will appear as a mixture of French and Altaians, even if the proportions may not be exactly correct.
I’ll get to this matter of proportions later on because it is relevant too.
What happens if we get the son of an Punjabi and Vietnamese and compare with French and Altaians? He will surely still show up as admixed. A simplistic conclusion might be that he is descendant from French and Altaians. This conclusion would be wrong, even if the confusion is understandable.
The Narmada hominin and “Neanderthal” admixture
Narmada skull
Thinking about this brought me (with some important help from readers – feedback is crucial) to the mysterious Narmada or Hathnora hominin (see here for an open access reference), the oldest of really big-brained humans and possibly a relative of Neanderthals (but not a true Neanderthal, among other reasons because they did not use Mousterian technology but Acheulean). The skeletal record of South Asia is quite scarce but this big-headed hominin is the last people we know about before African-like Middle Paleolithic technology appears c. 120,000 years ago (see here), probably with the first members of our species.
Yes, you read right: 120,000 years ago (more or less), the idea of a much more recent Out of Africa is almost certainly wrong, even if you will surely read such nonsenses for a while: the molecular clock pseudo-science cannot overrule archaeology.
It is at this moment uncertain whether the Narmada specimen and the probably much larger population it belonged to was a descendant of H. erectus or a descendant from H. heidelbergensis (and hence closely related to Neanderthals). Depending on which of these two options is correct, the scenario presented below will make sense or need to be revised.
I will consider, as suggested here by Michael Petraglia, that the Narmada specimen and related Indian population of the Early Paleolithic (which lasted until c. 100,000 years ago) were descendant of H. heidelbergensis, and hence cousins of Neanderthals. Why? Because they had Acheulean technology, which is associated with at least the late H. ergaster.
If this is correct, when we talk (after Green 2010) of Neanderthal admixture at low levels in non-African modern humans we may well be talking of admixture with anything within the broader Neanderthal family, in other words, with its ancestor H. heidelbergensis (cousin of our most direct ancestor H. rhodesiensis) and their descendants (Neanderthals and others, including probably the Narmada hominin and broader Acheulean-using population of South Asia.
A hypothesis strongly consistent with the coastal (or tropical) migration model
And I finally reach here to my hypothesis, to my explanation of the admixture episodes revealed by the Neanderthal Genome Project this eventful year of 2010. And I will do it with few words:
click to expand
The first admixture refers to the general non-African admixture with “Neanderthals”, which would actually have happened with their Indian cousins instead upon arrival to South Asia. This admixture would have affected all non-Africans, but as the case of the Karitiana (who only show some 0.9% of such admixture, much less than the rest) evidence, maybe not all populations exactly in the same amounts.
The second admixture refers to the specifically Melanesian hybridization with “Denisovans”, which would actually have been with their Indonesian pureblood relatives, H. erectus soloensis.
Makes sense? I think so. Of course, it is not set on stone but it seems a good hypothesis and should at least get some people chewing on this.
Special thanks for some key references to Terry T.  and Manju (but in general to all readers who take part in the discussions at the comments sections: keep the flow of ideas vibrant, please). I suspect that Terry will not like my conclusions because they end up in the coastal migration model that he hates so much. But well…
Appendix 1: the real apportion of Melanesian admixture with archaic hominins may be lower than suggested by Reich 2010.
They suggest (supp. info 8) that Melanesians would have as much as 7.4% of admixture with archaic species: 4.8% Denisovan plus 2.5% Neanderthal. But, if Denisovans are hybrids of H. erectus and H. neanderthalensis (as seems most likely, see above), then the real admixture with H. erectus would be an undetermined percentage but always less than 4.8%. As we know that the Neanderthal (or Heidelbergensis) component is 2.5%, it is most likely that the actual Erectus admixture in Melanesians is of only 2.3% or 2.4%, totaling 4.8%.
Appendix 2: very serious inconsistencies in the age estimates derived from nuclear DNA in Reich 2010.
In supp. info. 6, the authors provide data of genetic divergence between various modern populations, Neanderthals and Denisovans, expressed as fractions of the Homo-Pan divergence. They use the wrong time frame for this event (6.5 Ma) but even then the results make no sense.
Using a much more correct (according to modern best scientific understanding) of 8 million years, I get that the age of the migration Out of Africa would have happened between 650,000 and 500,000 years ago. The first figure is the distance between Yorubas and non-Africans and the latter the one between non-Africans.
This is a total nonsense (120,000 years makes sense, add some tens of thousands more if you wish but half a million years is simply not possible) and there must be a critical error somewhere. However I have not been able yet to discover what exactly is wrong. In any case, word of warning about accepting molecular clock age estimates in general and in particular when using nuclear (autosomal) DNA for this purpose.