May 19, 2016

35,000 year old mtDNA haplogroup U6 from Romania

I wouldn't be very surprised if many of the markers supposedly signifying recent gene flow Africa and Eurasia were actually quite old in Eurasia. The trouble is that reports of such gene flow were often based on simply observing that marker "X" occurs at a higher frequency in Africa than in Eurasia, so a common sense explanation is that it reflects limited recent gene flow between the continents. But, it is now known that common sense is not always the best guide, as e.g., ancient Europeans had mtDNA haplogroup M (in the past considered evidence of Asian admixture), Y-chromosome haplogroup C (ditto), and now U6.

The same should also apply to the Middle East where there has been admixture with Africans since the Islamic period at least. The existence of such admixture does not mean that every single lineage that occurs at low frequency in the Middle East and high frequency in Africa is diagnostic of this later period of admixture. Some of them could well be relics of old Middle Eastern populations. Who knows what people inhabited the presently inhospitable landscape of the Saharan-Arabian desert zone? The living populations can certainly make no claim to being the first ones there, but the genetic heritage of those earlier occupants may still persist in them in traces.

Similarly for the New World; in that case, there is a better case that European-looking lineages are indeed due to the colonization of the Americas over the last five centuries. However, that does not mean that all of them are, and we should be mindful of the possibility of pre-Columbian contact between the Old and New worlds.

Scientific Reports 6, Article number: 25501 (2016)

The mitogenome of a 35,000-year-old Homo sapiens from Europe supports a Palaeolithic back-migration to Africa

M. Hervella et al.

After the dispersal of modern humans (Homo sapiens) Out of Africa, hominins with a similar morphology to that of present-day humans initiated the gradual demographic expansion into Eurasia. The mitogenome (33-fold coverage) of the Peştera Muierii 1 individual (PM1) from Romania (35 ky cal BP) we present in this article corresponds fully to Homo sapiens, whilst exhibiting a mosaic of morphological features related to both modern humans and Neandertals. We have identified the PM1 mitogenome as a basal haplogroup U6*, not previously found in any ancient or present-day humans. The derived U6 haplotypes are predominantly found in present-day North-Western African populations. Concomitantly, those found in Europe have been attributed to recent gene-flow from North Africa. The presence of the basal haplogroup U6* in South East Europe (Romania) at 35 ky BP confirms a Eurasian origin of the U6 mitochondrial lineage. Consequently, we propose that the PM1 lineage is an offshoot to South East Europe that can be traced to the Early Upper Paleolithic back migration from Western Asia to North Africa, during which the U6 lineage diversified, until the emergence of the present-day U6 African lineages.

Link

8 comments:

capra internetensis said...

How the heck is Oase-1 within N, with N1 as a sister branch to an Oase1 + R + X clade? He's pre-N. And how is X2b a sister branch to R1b (R3) when they don't share a single mutation under N? How is an MN node maximally robust? Why is U6 supposed to be 49 000 years old when Muierii2 has one private mutation under U6, Ust' Ishim man has 1 private mutation under R, U6 is 5 mutations under R, and Ust' Ishim is 45 000 years old?

Krefter said...

" The trouble is that reports of such gene flow were often based on simply observing that marker "X" occurs at a higher frequency in Africa than in Eurasia, so a common sense explanation is that it reflects limited recent gene flow between the continents. But, it is now known that common sense is not always the best guide, as e.g., ancient Europeans had mtDNA haplogroup M (in the past considered evidence of Asian admixture), Y-chromosome haplogroup C (ditto), and now U6."

Exactly. Ancient European DNA has turned European genetics upside down. Most of what we've learned wasn't predicted using only modern DNA. The same will be true for the rest of the world. There could have been very interesting people living in the Congo or Arabia or India 1,000s of years ago that went extinct.

DDeden said...

"...pre-Columbian contact between the Old and New worlds."

https://www.academia.edu/25063114/2008_Storey_et_al._Pre-Columbian_chickens_dates_isotopes_and_mtDNA._PNAS

eurologist said...

Here we have it again (make another note to the history of science): "The time of the most recent common ancestor (TMRCA) for U6 was estimated to 35.3 kya BP" - while "Our estimates of the haplogroup U6 TMRCA that incorporate ancient genomes (including PM1) set the formation of the U6 lineage back to 49.6 ky BP."

Another ancient specimen, another earlier date for a lineage. In this case a whopping 40%! I think we have another data point demonstrating that the mutation rate is non-linear, as is expected using common sense.

Anonymous said...

Out of Africa itself relies ONLY on this same kind of coincidental circumstancial evidence as evidence. These "back migrations" just keep adding up. Hopefully we can throw this whole theory in the garbage can soon, where it belongs.

capra internetensis said...

@eurologist

Why is the estimate from this paper to be preferred to previous ones? The recent paper on Ice Age mitogenomes (Posth et al) used a lot more Palaeolithic samples and gave much younger ages. It also had a phylogeny that made sense.

The new sample isn't responsible for the difference, because it fits perfectly into the previous tree. PM1 sample is U6 with one private mutation, so 6 mutations below R, and dated to ~35 000 cal ybp. Ust' Ishim man has one private mutation below R and is dated to ~45 000 cal ybp. The Posth et al paper gives a central estimate of 2.74E-8 mutations/site/year. 5 mutations below Ust' Ishim would be ~11 000 years, or ~34 000 years ago for PM1, which is dead on, and would make U6 itself ~36 000 years old, in agreement with the previous estimate.

Meanwhile Hervella et al's new estimate of 49.6 kya for U6 and ~61 kya for R means that 5 mutations happened in 11 000 years leading to U6, then one mutation occurred in 15 000 years after that going to PM1. Possible, yes. Parsimonious, no. Non-linear? Doesn't show up in comparison between actual dated samples.

eurologist said...

capra,

That's not how the statistics of a low number of mutations works. You can still find examples of haplogroups way up the line, in today's populations.

We should acknowledge that pretty much all modern literature points to a non-linear effective mutation rate - although the reasons are still debated and murky. From the first estimates researchers found a huge discrepancy of nearly an order of magnitude between pedigree and phylogeny based rates. To me, the most important reason is the simplest: many mutations are at least somewhat deleterious, and in accumulation, more so. That is, most lines with higher mutational load won't make it, whereas lines with low load (a smaller number of mutations) will prosper - making the effective mutation rate much slower the farther you look back. And indeed, that is exactly what is observed. The effective mutation rate is likely roughly constant (and quite low) for a period of a few million ya, then accelerates slowly to about 200,000 ya, and then has accelerated at a faster pace ever since - even if assuming a neutral background.

See also:
Characterizing the Time Dependency of Human Mitochondrial DNA Mutation Rate Estimates
B. M. Henn et al, 2008;
Molecular Biology and Evolution Volume 26, Issue 1, Pp. 217-230.
http://mbe.oxfordjournals.org/content/26/1/217.full.pdf+html

(specifically for mtDNA).

capra internetensis said...

@eurologist

I know that the mutations do not accumulate at a steady rate, and I know there's evidence for time dependency and long-term changes in mutation rate. Yes, any given sample could have a number of mutations far below average for its time. My point was that this one has the expected number of mutations, so if you plugged it into the tree from the previous study, which gave a much younger TMRCA for N and R, it would not be expected to change the timeline significantly, so the older date of the more recent study was unlikely to result from the new ancient DNA evidence in itself.

Compare the dates and branch lengths of the Early Palaeolithic samples we have. They are consistent with quite recent TMRCAs for N and R.