Theory One: Ancient African Population Structure
It could be due to a shared common ancestor, Homo heidelbergensis, which could have split into Neanderthal population in the North and early modern humans in Africa around 350,000 to 500,000 years ago. In this scenario the differences in the amount of apparent Neanderthal admixture between non-Africans and Africans was due to population structure in Africa, with the Out of Africa population drawing on a different part of the genetic diversity of modern humans at the time than the population which in ancestral to most Africans today.
Keep in mind that much of the population diversity present at the Out of Africa moment in Africa was found in populations that subsequently left no descendants or are survived only by relic populations like the Pygmies and Khoisan today, or only have ancestors who slightly introgressed into other, more successful African populations.
Different African and Eurpean hominid populations around 100,000 years ago could have had different levels of relatedness to each other, presumably on a North to South cline, with the populations closest to Europe being most similar to Neanderthals and the populations most distant from Europe being least similar. If many of the intermediate hominin populations descended from Homo heidelbergensis between Northeastern African modern humans who were necessarily the ancestors of non-Africans on one hand, and the Southern African or Western African modern humans, on the other, subsequently went extinct, the genetic trace that would be left in modern humans would look a lot like an admixture model, but without the hot interspecies sex. Likewise, hominin populations intermediate between Northeastern African modern humans (pre-Out of Africa), and Neanderthals, in the Levant, would also have left no genetic traces. Fans of Jean M. Auel's Clan of the Cave Bear series wouldn't be vindicated after all.
This is the essence of the argument made by Andrea Manica and Anders Eriksson in a paper in the Proceedings of the National Academy of Sciences, which was acknowledged by viewed a less parsimonious in the seminal Neanderthal admixture paper. (Archaeogenetics blogger Dienekes is a fan of this theory.)
Theory Two: Neanderthal Admixture
Or, it could it be that the sexier theory that Homo neanderthalensis interbred in small numbers with some of the earliest ancestors of all non-African modern humans is correct?
This is the proposal of Svante Pääbo "who led the sequencing of the Neanderthal genome in 2010 and has championed the idea that modern humans interbred with Neanderthals[.]"
Pääbo has co-authored a paper, which is yet to undergo peer-review, to further support his thesis that humans and Neanderthals did in fact interbreed. "We find that the last gene flow from Neanderthals (or their relatives) into Europeans likely occurred 37,000-86,000 years before the present, and most likely 47,000-65,000 years ago," he writes. "This supports the recent interbreeding hypothesis, and suggests that interbreeding may have occurred when modern humans carrying Upper Paleolithic technologies encountered Neanderthals as they expanded out of Africa."These Admixture Dates, If Recalibrated, Fit An Early Out of Africa In The Levant That Lived
If the dates are calibrated to fit archaeological suggestions of date that make sense, to be longer by a factor of 1.6, as I have advocated in previous posts (the difference mostly reflects a probably inaccurate mutation rate used to develop the estimate of the number of years ago that genetic populations became distinct), one gets a date range of 75,200 to 104,000 years ago.
This neatly coincides with the earliest time period in which there is archaeological evidence of Neanderthal and modern human co-existence in the Levant, after which modern humans reappear in the Levant about 50,000 years ago.
The adjusted date range would be powerful evidence that the first Out of Africa period in the Levant was not "Out of Africa that failed", but instead, was the source of the population that is the dominant ancestral population for all non-African populations today. In this scenario, rather than dying out, after which Eurasia was recolonized by a second wave of modern humans leaving Africa, the early Out of African Levantine population was dispersed around 75,000 years ago from the Levant to refugia (probably in a wetter interior of Arabia, in the less inundated Persian Gulf, in Iran, and/or in South Asia) that later rebounded and ultimately populated the rest of the Earth.
Why Don't Europeans Have More Neanderthal Admixture Than Asians?
How can one explain the similarity in levels of African admixture between Europe and Asia in an admixture theory, despite the fact that Neanderthals co-existed with Cro-Magnon modern humans in Europe for many thousands of years, but in Southeast Asia and East Asia for far shorter time periods?
Did the propensity of modern humans to interbreed with Neanderthals decline in the Upper Paleolithic? I don't think that it did.
The more likely scenario, I think, is that Neanderthal co-existence with modern humans in the Middle East ended far sooner than it did in Europe, and that most of that ancestry of Europeans today comes not from the Cro-Magnon population that had a prolonged period when there could have been admixture, but from Middle Eastern populations with less Neanderthal exposue. These peoples repopulated Europe after the Last Glacial Maximum (around 20,000 years ago, long after the Neanderthals had gone extinct) and in subsequent waves of population expansion around the beginning of the Neolithic Revolution in Europe (plus or minus a few thousand years) and with Indo-European language expansion. This would dilute the amount of excess Neanderthal admixture in Europe.
Since the level of Neanderthal admixture would have been modest even in populations with long periods of co-existence with Neanderthals relative to the non-African baseline, even in more admixed populations, it could be that these elevated Neanderthal admixture levels from Cro-Magnon populations are almost indistinguishable from background random noise in modern populations.
For example, suppose that later waves of populations from the Middle East provided 80% of the autosomal genetic source for modern Europeans (one number that has been seriously proposed based on assumptions made about uniparental genetic markers) and 20% of the autosomal genetic source for modern Europeans was Cro-Magnon. Then, assume that the Middle Easterners were 4% Neanderthal admixed and that the Cro-Magnon were 8% Neanderthal admixed (half due to their remote Middle Eastern source ca. 43,000 years ago, and the other half from further admixture in Europe). The end result would be a modern European population with 4.8% Neanderthal admixture on average, a percentage that has a standard deviation in Europe on the order of 0.5 percentage points in the data so far used to derive that figure.
More realistic numbers are something like this:
* The level of Neanderthal admixture may be somewhat higher in the minority of modern European populations, probably not more than 10% of all Europeans overall (mostly in the far Northeastern part of Europe), with the largest genetic contributions from the last European hunter-gatherers to convert to herding and farming, who might have, perhaps as much as 20%-30% of ancestry in these populations traceable to Cro-Magnons.
* Don't forget that post-last glacial maximum relict European hunter-gatherer populations, even before the demic impact of Neolithic and metal age migrants, would not have been 100% Cro-Magnon. Almost all of Northern Europe was 99.99% depopulated of modern humans because it was under glaciers (there may have been a warm protected valley or current warmed coastal area in a few isolated spots that could support a village or two), leaving relict populations mostly in Iberia, Italy and far Southeastern Europe. When the glaciers retreated, some of the repopulation probably had its source not from these Cro-Magnon refugia, but from the Middle East. By the time that the herders and farmers arrived, these source populations would have had at least several thousand years to reach something close to
* In contrast, most European populations, who make up perhaps 90% of all Europeans, might have on average 10% Cro-Magnon ancestry or less.
* The level of Neanderthal admixture in Cro-Magnon populations might have been 50% more than in Middle Easterners (i.e. Cro-Magnon populations might have been 6% Neanderthal), rather than twice as great.
* This would suggest an expected average level of European Neanderthal admixture of 4.23% relative to a baseline of 4% +/- circa 0.5%, and the existence of population structure in Europe would mean that the choice of sample individuals from which you made the estimate could skew the sample far more than the excess of the European average over the shared non-African baseline percentage of Neanderthal admixture. Thus, the excess Neanderthal admixture would be almost invisible.
Why Is There No Neanderthal mtDNA or Y-DNA in modern humans?
There are also plausible models that can account for the absence of Neanderthal mtDNA or Y-DNA in any modern humans.
In these models, a principle called Haldane's law insures that almost all of the fertile Neanderthal-modern human hybrids were probably female in the first generation, accounting for the lack of Neanderthal Y-DNA. Haldane's law says that mixed sex chromosome interspecies hybrids (XY males in the case of hominins) tend not to be born, or to be infertile if they are born, relative to homogeneous sex chromosome interspecies hybrids (XX females in the case of hominins).
Neanderthal mtDNA's absence is explained by the theory that hybrid children grew up with a tribe made up of the species of the mothers rather than the fathers, and that the hybrid children born into Neanderthal tribes died out with the rest of the Neanderthal population, while the hybrid children born into modern human tribes have descendants who are alive today. In this scenario, there could have been just as many hybrid children with Neanderthal mothers as there were with modern human mothers, and yet have no Neanderthal mtDNA in modern humans.
Of course, in both the Y-DNA and mtDNA cases, isolated exceptions to the rule could wash out of the gene pools due to genetic drift.