Monday, January 26, 2015

"Ancient North Eurasians": North-South Ancient Admixture in Central-South Eurasia

Back in February of 2014, I put up a post discussing an Admixture components pattern that appears across populations extending from Russia in the Chuvash, to the North Kannadi people in Southern India. I describe this pattern as the "Northeast Asian" - "East Asian" admixture components pattern.
This pattern can also be observed in the Admixture run shown in the paper The genome of a Late Pleistocene human from a Clovis burial site in western Montana, Rasmussen et al, 2014, Extended data, Figure 3:

Pronounced "Northeast Asian" like components can be observed in the Mari (related to the Chuvash), as well as Northeast Asian populations.
Pronounced "East Asian" like components can be observed in South Asian populations such as the North Kannadi people, as well as in populations of Southeast Asia.
These two components seem to be dispersed into each other from an ancient date, as described in the February 9th post.
I have seen various blogs and papers discussing the presence of an "Ancient North Eurasian (ANE)" component.  In them, it is proposed that this "ANE" component is the cause for the "eastward" population shift in Europeans since the Mesolithic. 
I would propose that this "ANE" component is not a single "component", and does not come from a single geographic location in the last 20,000 years or so.  It is widely dispersed.  Therefore, it is likely a very ancient population process which is derived in part from archaic admixture.
In Europeans, the source for this "ANE" component likely is not derived from a single source or entry point into Europe.  If the Motala12 ancient Mesolithic DNA sample from Sweden is any indication, "ANE" like components appear to have been diffusing into Europe since the Mesolithic.  These "ANE" diffusion processes continue to the present day.
Finally, apart from the fact that the Mal'ta MA1 sample from Lake Baikal, shown on the PCA of the Lazarides paper 2013, happens to line up on the PCA with populations such as the Chuvash, I strongly doubt that the R1* y-DNA haplogroup is widely correlated with "ANE" like components.  The Mal'ta MA1 population was probably on the leading edge of an eastward expansion that was not shared across the R1* extended family. 
[ Author's note (5/1/2015):  By the statement "I strongly doubt that the R1* y-DNA haplogroup is widely correlated with "ANE" like components", I specifically mean that there is not a strong correlation of the R1b ydna ghs.  However, there does appear to a stronger correlation of R1a ydna hgs with the "ANE" admixture component.  The reason for this split in the correlation of R1a and R1b ydna hgs is likely because these groups split prior to the last Ice Age, with R1a positioned on the Steppe 25,000 years ago, and R1b positioned somewhere in Europe 25,000 year ago. ]
Marnie Dunsmore (January 26th, 2015)
The February 9th, 2014 post:
Quite early in the history of this blog, on November 3rd, 2010, I wrote a post discussing a minor "signal" I had noticed buried in the Dodecad admixture results for populations of Eastern Europe, the Middle East and South Asia.  The "signal" consisted of a combination of two Admixture components: "Northeast Asian" and "East Asian".  You can read about that early post here.  I've added a few thoughts on this topic recently, noting that the pattern is likely evidence of ancient bidirectional gene flow. (See Max Born on receiving the 1954 Nobel Prize . . ., Update, February 1, 2014.) 
In this post, I look at additional major components such as the "West Asian" component.
Reviewing observations from the previous posts, a number of deductions are possible:
1.  The "Northeast Asian"-"East Asian" admixture components are geographically correlated.
2.  The pattern extends from the Volga region of Southern Russia (Chuvash), into the Caucasus and Black Sea,  then into Pakistan, splitting westward through Syria into the Fertile Crescent, touching even into Egypt, and separately, trailing south through Pakistan extending west along the South Asian coast.
3.  The pattern is widely diffused compared to other admixture components.  In particular, it is much more widely diffused compared to the "North European" component and the "West Asian" component.  This wide diffusion is possibly indicative of great age since coalescence.
4.  As argued in the February 1st update of the post Max Born on receiving the 1954 Nobel Prize . . ., the "North European" component is likely correlated with pre-Ice Age hunter gatherers. 
5.  The "North European" component does not appear to be correlated with the "Northeast Asian"-"East Asian" pattern and is likely a superposition on it.
6.  The "Northeast Asian"-"East Asian" component does not appear to be present in the following populations considered in the Dodecad Admixture run:  Lithuanians, Cypriots, Spanish, French Basques, Tuscans, French, Northern Italians, Armenians, and Mozabite Berbers.  It is therefore uncorrelated with admixture in these populations.
In this post today, I look at three other major Admixture components with respect to the "Northeast Asian"-"East Asian" pattern.  These are the "West Asian", the "Southwest Asian" and the "Southern European" components.

Again, the data is normalized on the "Northeast Asian"-"East Asian" pattern. The remaining un-normalized components are allowed to float above this.  Figure 1 shows the zoomed in view with the normalization pattern running along the bottom.  This normalization approach corresponds with a North-South geographic pattern, as described in item "2." above.  Figure 2 shows the same data, but from a "zoomed out" perspective.  Populations are listed below Figure 2.

Figure 1:  Admixture components with normalized "Northeast Asian"-"East Asian" pattern show at bottom.  Components are arranged with highest "Northeast Asian" contribution on the left and highest "East Asian" contribution on the right.
Figure 2:  As in Figure 1, but "zoomed out" to fully show the un-normalized major components floating above the normalized "Northeast Asian"-"East Asian" pattern.
    1:  Chuvash
    2:  Lezgin
    3:  Georgian
    4:  Sindhi
    5:  Belorussian
    6:  Adygei
    7.  Pathan
    8:  Syrian
    9:  Turk
   10: Jordanian
   11: Romanian
   12: Ashkenazi
   13: North Kannadi
   14: Gujarati
   15: Uygur
   16: Burusho
   17: Egyptian

Looking at the above plots, the essential question to be asked is this:  Is there a correlation between the "Northeast Asian"-"East Asian" pattern and
1.  the "West Asian" component,
2.  the "Southwest Asian" component, or
3.  the "Southern European" component
The "West Asian" component is widely distributed.  In a few noticeable cases, such as the Chuvash, the Uygurs, the Burusho, and the North Kannadi, the "West Asian" component is only a very low level admixture component.  Based on this, it cannot be argued that the "West Asian" component is strongly correlated across the ancient "Northeast Asian"-"East Asian" pattern.  Moreover, the "West Asian" component is likely a superposition on this pre-existing population.  On the other hand, it has already been argued in another post that the point of coalescence of the "West Asian" component is somewhere in the Caucasus.   The distribution of the "West Asian" component centered about Georgians is vaguely apparent here.
The distributions of the other two major components "Southwest Asian" and "Southern European" are sparse and loosely distributed.   There does not appear to be an easily observable correlation between the ancient "Northeast Asian"-"East Asian" pattern and these components, at least from this data.

No comments:

Post a Comment

Comments are enabled.