Razib Khan takes note of a New York Times review of a new book on the arrival of modern humans to the Americas incorporating new developments, essentially all of which have previously been blogged here. I left a couple of extended comment which I will repeat here with some minor edits (not as a block quote since I am quoting myself after the quotations he mades from the story below). The story is entitled "Did the First Americans Arrive via Land Bridge? This Geneticist Says No" and he quotes from it, this passage.
Raff skillfully reveals how well-dated archaeological sites, including recently announced 22,000-year-old human footprints from White Sands, N.M., are at odds with the Clovis first hypothesis. She builds a persuasive case with both archaeological and genetic evidence that the path to the Americas was coastal (the Kelp Highway hypothesis) rather than inland, and that Beringia was not a bridge but a homeland — twice the size of Texas — inhabited for millenniums by the ancestors of the First Peoples of the Americas.
As is often the case, I find this story to be overhyped (as the author, no doubt hoping for a best seller hopes it will be).
The linked New York Times coverage strongly implying that Clovis First is the paradigm was disappointing for putting up a straw man. Clovis First hasn’t been mainstream for many decades.
The important of pre-glacial melt migration, while just recently firmly established with credible evidence, is also overstated. The genetics and archaeology suggest that the pre-glacial melt, pre-Last Glacial Maximum population of modern humans in the Americas had almost no distinguishable demographic impact on a percentage basis, and also had negligible ecological impact and left few archaeological relics (in addition to a complete lack of human remains discovered so far after quite diligent study).
We are talking migrations in the low hundreds in that time frame, who mostly either died out or were demographically overwhelmed by the main Founding population of the Americas. (By the way, I like Greg Cochran’s description this population as “the Predecessors” at his West Hunter blog linked in the sidebar).
Even if Paleo-Asian ancestry in South America is from the glacial or pre-glacial era (and there are good, almost irrefutable, reasons rooted in population genetics and statistics to conclude the the Paleo-Asian ancestry seen in South America can’t have anything remotely close to that time depth), the percentage of the aggregate indigenous North, Central and South American gene pools attributable to Paleo-Asian ancestry is tiny (on the order of parts per 100,000, or parts per million, or less, for a population that should benefit from a Founder effect to the extent that it was at all genetically distinguishable from the Founding population and hadn’t either collapsed before the Founding population wave arrived, or been almost completely replaced without much introgression at all by the waves of migration for the Founders).
Likewise, a coastal migration along the coast of Beringia and the Pacific Coast of North America is not all that materially different from a land bridge with a long Beringian stand-still which is the current hypothesis. No one is suggesting that people could have made a maritime trip from Northeast Asia to North America without the existence of the Beringian land bridge even if they coast hopped the trip in boats that largely stayed within sight of shore, instead of walked the distance.
The biggest really recent development that does depart from the paradigm but isn’t getting top billing in the popular press is the strong suggestion supported by genetics and archaeological evidence (and even linguistic evidence to some extent), that there were two separate waves of migration from Northeast Asia and Beringia to North America that were fairly close in time (maybe 500-2000 years apart), around the conventional pre-Clovis date for the Founding population of the Americas. The suggestion is that there was one wave of Founders largely confined to North America that is more Northeast Asian, that eventually admixed almost completely with the population from the other wave, and the other more Beringian wave being the only wave that reached Central and South America. These waves would have been similar in population genetics, but distinct enough to still leave traces in uniparental haplogroup frequencies and presence in modern populations of indigenous Americans.
The new developments also don’t alter the paradigm with respect to Paleo-Eskimo, Na-Dene, and proto-Inuit populations much later in time (basically Bronze Age and later).
These new developments also don’t disturb the conclusion that the Solutrean hypothesis (i.e. that a significant share of North Americans migrated there as a founding population by sea from Europe at around the Founding population era) is false.
Also, I take issue with the way the New York Times article describes the statement that:
She eviscerates claims of “lost civilizations” founded on the racist assumption that Indigenous people weren’t sophisticated enough to construct large, animal-shaped or pyramidal mounds and therefore couldn’t have been the first people on the continent.
In fact, the Americas have a number of "lost civilizations" of collapsed civilizations, albeit home grown and not due to European or Asian influences or to "ancient aliens".
There was a Neolithic civilization in the Amazon that collapsed. The Cahokian culture was an empire that rose and fell, as did the empires of the Mayans, the Aztecs, the glyph building culture of the South American highlands, and the Ancient Puebloan culture, for example.
There is nothing racist about the conclusion that these relatively sophisticated indigenous cultures reached a point from which there was technological and cultural regression for a long period of time causing those cultures to be "lost civilizations" in a fair sense of that phrase.
Uh. The Paleo-Asian ancestry is much higher than just a few parts in 100,000.
ReplyDelete@Ryan
ReplyDeleteIn particular individuals, yes, it is present at the lowish single digit percent level, but the percentage is out of the indigenous American gene pool for all of the Americas as of 1492 CE.
It is found only in a dozen or so Amazonian basin tribes all of which have small populations, and is entirely absent in North America, Central America, and most South American populations.
Suppose that the average percentage of Paleo-Asian ancestry in someone who has it is 2% (which is high). Suppose further that there are 6,000,000 full blood equivalent individuals (treating, for example, someone with 25% indigenous American ancestry as 0.25 FBEs) in the Americas today (which is the right order of magnitude) and that there are 60,000 people with Paleo-Asian ancestry.
So, there would be 1,200 FBEs of Paleo-Asian ancestry, out of 6,000,000 FBEs of indigenous American ancestry, which would be 20 parts per 100,000 of Paleo-Asian ancestry in the indigenous American gene pool, a back of napkin estimate that is probably high.
Notes to self re Paleo-Asian genetics math.
ReplyDelete* No Paleo-Asian Y-DNA or mtDNA. This suggests that Paleo-Asians were a low percentage of the gene pools initially because uniparental markers are lost in an expanding population like the one in the New World only if it is a low percentage and the losses of uniparental markers happens in the early generations. It is also easier to lose both Y-DNA and mtDNA if the initial population of Paleo-Asians was either male dominated or female dominated, rather than gender balanced.
* We know what low percentage introgression of a predecessor population a long time in the past looks like. It looks like Neanderthal and Denisovan ancestry, which is close to uniform. And we know from models that most of the homogenization of archaic ancestry percentages happens in the first less than 100 generations.
* There is great intra-population variation in Paleo-Asian ancestry, which implies recent dispersal across the Amazon basin tribes. It is the kind of variation we see in Uralic ancestry in Hungarians or Mongolian ancestry in Koreans, perhaps. Whenever it arrived in the Americas, Paleo-Asian ancestry had to disperse early (but it has to be old enough to produce the low percentages in the people who have it, there is nobody with 25% Paleo-Asian ancestry, so the last full blooded Paleo-Asian ancestor has to be not less than about 5-8 generations ago (150-240 years ago), although it could be much older.
* Suppose Paleo-Asian ancestry from the Founding population era or older introgresses at a small percentage of the community at the time of contact and that population remains isolated until the much more recent time period of dispersal in the Amazon basin tribes. Then, you've got a small population that is genetically isolated for more than 10,000 years. Genetic drift in that isolated population from other New World people would stick out like a sore thumb as it does in the Kalash people. But that isn't present.
I can't see any scenario in which the distribution of Paleo-Asian ancestry in the Americas could arise from a source that is all that old. Probably not even back to 3000 BCE on the oldest end of when Paleo-Eskimo populations could have arrived in North America.
Instead, I suspect that Predecessor population genetics had undiscernable impact because the Predecessors were probably close genetic cousins of the Founding population with little genetic drift separating them and were very small number relative to the Founding population at the time of first contact, so a combination of dilution and similarity made them basically invisible in modern population genetics and available ancient DNA.
This still doesn't solve exactly what Paleo-Asian DNA is, but we can infer enough from what we know to rule out lots of possibilities.
Closer to 120 million FBEs of indigenous ancestry in the Americas as a whole. See, e.g., https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1005602
ReplyDeleteThis would bring my estimate (still high) to one part per 100,000 in the overall gene pool. So, parts per million is in the right ballpark.
"No Paleo-Asian Y-DNA or mtDNA. "
ReplyDeleteAssuming the Pealo-Asian Y-DNA wasn't Q though and the mtDNA wasn't B (B shows a coastal dispersion pattern if memory serves).
"Closer to 120 million FBEs of indigenous ancestry in the Americas as a whole. "
We have only a handful of unadmixed samples to go from though, not 120 million.
My thinking at least is that there was never a "pure" Paleo-Asian population in the Americas... just a population with a higher degree of admixture from that source.
ReplyDeleteThe Yana paper already modelled Yana as being part Paleo-Asian (ie Onge-like) so I think there's a baseline there that came with ANE to all populations that have ANE (including Europeans). The signal we see is just a matter of degree.
The Paleo-Asian Y-chromosome groups seem to be K2b, C and D in my view. So Q would have come from ANE's Paleo-Asian side anyways.
Re uniparental markers
ReplyDeleteNeither Y-DNA haplogroup Q nor any of the mtDNA haplogroups found in South America are found in combination the Paleo-Asian type autosomal genetic profiles anywhere else in the world. You don't see Paleo-Asian in other ANE rich populations either - it is a Southern Asian profile. North Asians and Japanese people with Y-DNA D don't have autosomal Paleo-Asian ancestry.
"We have only a handful of unadmixed samples to go from though, not 120 million."
We have lots and lots of New World modern and ancient DNA samples that were selected with an eye towards capturing samples from all subpopulations in a structured population of about 400 million people.
Due to weak panmixia assumptions, and the fact that Paleo-Asian ancestry involves thousands of SNPs only a modest portion of which need to be present to be detectable, just a handful of samples from anyone in a population is a powerful predictor of whether anyone in that population (regardless of its census size) has a particular kind of ancestry, particularly in the case of ancestry that introgressed into the gene pool of one of the main contributors to the current gene pool thousands of years ago. You shouldn't be able to miss the presence of Paleo-Asian ancestry that introgressed into the gene pool more than 5000 years ago in the Mexican gene pool, for example, in a sample of 65 Mexicans with indigenous ancestry even through there are 129 million people in Mexico, and far more than 65 million of them have some indigenous New World ancestry. And, pretty much all large New World populations have that many modern samples.
There s no Paleo-Asian ancestry in any of the New World populations with large census sizes. It is found only in Amazonian basin populations with tiny census sizes, and only at low percentage of Paleo-Asian ancestry in the individuals in which it is present.
Paleo-Asian ancestry were present in more than one or two members of the Founding population from 14000 to 15000 years ago in any specific continental or subcontinental region of the Americas, if it survived at all, it would be ubiquitous, unless this ancestry codes has functional traits that negative selective fitness effects, as a powerful Founder effect.
To get the frequencies you see, there really couldn't have been more than one nuclear family with this ancestry in the founding population of the Amazonian basin and they couldn't have any descendants in any other part of the Americas. But, in that scenario, you'd either have a highly uniform percentage of Paleo-Asian ancestry among all people with indigenous ancestry in the Amazonian basin (which you don't) or you'd have a highly drifted and divergent subpopulation that didn't admix with other South Americans for more than 10,000 years which would stick out like a sore thumb.
Even within those tiny Amazonian tribes, there is great variation in proportions of Paleo-Asian ancestry which demonstrates that the dispersal of this ancestry to these small tribal populations (either through spouse exchange or tribal mergers between tribes with and without it) is not all that many generations old (dozens of generations at most and certainly less than 2000 years).
Anyway, the bottom line is that Paleo-Asian ancestry really does make up something on the order of parts per 100,000 to parts per million of the overall indigenous autosomal gene pool of the Americas.
"Neither Y-DNA haplogroup Q nor any of the mtDNA haplogroups found in South America are found in combination the Paleo-Asian type autosomal genetic profiles anywhere else in the world."
ReplyDeleteYou find close relatives though. Like in the Aeta.
@Ryan
ReplyDeleteDo you have any data on the uniparental genetics of the Aeta?
https://en.wikipedia.org/wiki/Aeta_people
I think you miss the point of the 'lost civilizations' quote, probably due to the scare quotes used. Let me re-jigger the sentence for you.
ReplyDelete[She eviscerates the racist "lost civilizations" trope that Indigenous people weren’t sophisticated enough to construct large, animal-shaped or pyramidal mounds and therefore couldn’t have been the first people on the continent.]
So in fact, the sentence does refer to the 'Atlanteans/Egyptians/Aliens must have done it' genre of fantasy, and not real history. And the fact that the NY Times needs to debunk bunk says more about virtue signalling than it does the study of prehistory.