PDA

View Full Version : South Asian results in 23andme



Skyfall
09-28-2015, 10:20 PM
I understand that 23andme has a "South Asian" component when displaying ancestry results; However, despite this, I have seen results in the past where the breakdown of ancestry has been something like this:

http://i.imgur.com/jVT8gxb.jpg

This is the chromosome painting of an "Indian man" and it says that 90% of the ancestry is from Europe, and 10% from Asia. From talking to other members on 23andme who are South Asian and have gotten their results back, it seems that most South Asians are scoring in the range of 10-35% Asian, as Razib Khan points out in this post: http://genomesunzipped.org/2011/02/guest-post-by-razib-khan-my-personal-genome.php

Razib states that most South Asians are 75% Caucasian and 25% Asian according to the results generated by 23andme. He also states that this "Asian" value fluctuates depending on region and caste, with it reaching a low of 10% Asian in the Northwest of the subcontinent and reaching a high of 35% in the East and the South. He also cites his own results as a Bangladeshi, stating that he is 57% Caucasian/European and 43% Asian, a result that is highly irregular and a departure from the norm for most South Asians, who are, on average, 25% Asian.

Does this mean that the average Indian is 75% European/Caucasian and 25% Asian, with some being more Caucasian, particularly in the Northwest upper castes, where the upper castes can be up to 90-95% Caucasian, and some being less Caucasian, particularly in the East and South, where the lower castes can be ~60% Caucasian? Even the National Geographic Geno 2.0 project substantiates these numbers to some extent, stating that the average NW South Asian is 75% Caucasian and 25% SE Asian.

Are these values correct? It seems like there is a lot of confusion regarding the ancestral components of South Asian populations, and I'd appreciate some insight into this matter. Thanks a bunch!

Varun R
09-29-2015, 12:33 AM
Indian average likely closer to the 30-35% E Eurasian range. Not all of this is ASI, as has been stated by Bored.

Skyfall
09-29-2015, 01:02 AM
How much of the 30-35% would be non-West Eurasian or ASI/East Asian? 50%? And would this mean that the average Indian is also 65-70% Western Eurasian? Again, thanks for the insights.

Varun R
09-29-2015, 01:19 AM
All of the E Eurasian is, well, E Eurasian. Some of the remaining W Eurasian is not part of "Crown Eurasian" but is rather Basal Eurasian. There is a thread on "what is basal eurasian" that you may find inciteful.

Skyfall
09-29-2015, 01:46 AM
OK I see. So the East Eurasian ancestry is completely East Eurasian, but the West Eurasian ancestry has a "basal Eurasian" component? Quick question: is the "basal Eurasian" component something that makes the overall ancestry less West Eurasian, or is it simply a different type of West Eurasian ancestry? I would think the "basal Eurasian" is also completely West Eurasian in composition due to this piece of evidence:

"Basal Eurasian is obviously Caucasoid or Caucasoid-like because Stuttgart is almost half Basal Eurasian but has 100% West Eurasian ancestry."

Of course the Stuttgart is referring to this:

"The ancient samples appear to be mostly West Eurasian in their ancestry, although the hunter-gatherers are also inferred to have greater or lesser extents of the eastern non-African (ENA) component which is lacking in Stuttgart [Neolithic farmers]. This is consistent with the positive f4(ENA, Chimp; Hunter Gatherer, Stuttgart) statistic reported in SI12, which we interpret there as showing that ENA populations are closer to European Hunter-Gatherers than to Stuttgart. [...] MA1 shows some affinity to [the south Eurasian component maximized in Papuans], in contrast to more recent Eurasian hunter-gatherers who continue to mainly show ties to Native Americans."

"So we can see that Early European Farmers (EEF) represented by Stuttgart had 0% admixture, West European Hunter-Gatherers (WHG) represented by Loschbour and Motala had an average of ~10% admixture, and Ancient North Eurasians (ANE) represented by MA1 had ~45% admixture."

6088

And perhaps this is why initial studies lumped ASI with the Onge:

"Another big issue for this model is simple, in truth. Australo-Melanesians, Andamanese and East Asians are seemingly closer to each other than they either is to Europeans often prompting various population geneticists to lump them into a sort of "clade" with one another like "Eastern Non-African" in Lazaridis et al. 2013 or similarly in Kay Prüfer et al. 2014 [5] as I touch upon here."

As stated at this link: http://anthromadness.blogspot.com/2015/09/is-eastern-non-african-situation-more.html

Essentially, since they are in the same clade, ASI was erroneously assumed to be = to Onge, when in reality, it is East Asian/SE Asian in ancestry.

Thanks again for your insights!

Varun R
09-29-2015, 02:00 AM
Basal Eurasian is not part of the Crown Eurasian group. It was christened in order to explain why Loschbour (pure WHG) was closer to Han (ENA) relative to modern Europeans than would otherwise have been anticipated. Calling it West Eurasian is somewhat of a misnomer, although it is widespread in W Eurasian dominated regions today.
I would read that thread for more info.

Megalophias
09-29-2015, 02:18 AM
http://www.anthrogenica.com/showthread.php?5366-What-is-Basal-Eurasian (What is Basal Eurasian?)

Skyfall
09-30-2015, 05:11 AM
So I just wanted to recap here, when Razib stated that most South Asians were scoring in the range of 10-35% Asian in his above, was he erroneously underestimating the actual East Asian ancestry in South Asians, or was that simply a function of the sample that was present on 23andme?

Again, Razib states that the average South Asian is 75% Western Eurasian and 25% Eastern Asian according to the results generated by 23andme. He also states that this "Asian" value fluctuates depending on region and caste, with it reaching a low of 10% Asian in the Northwest of the subcontinent and reaching a high of 35% in the East and the South. He also cites his own results as a Bangladeshi, stating that he is 57% Western Eurasian/European and 43% Asian, a result that is highly irregular and a departure from the norm for most South Asians, who are, on average, 25% Asian.

I noticed that you stated the average value for "Asian" is more in line with 30-35% East Asian, so are you essentially stating that Razib was wrong to extrapolate the sample on 23andme to all subcontinental individuals? This seems plausible to me, again only because of the smaller sample size.

Does this mean that the average Indian is 70% Western Eurasian and 30% Eastern Asian, with some being more Western Eurasian, particularly in the Northwest upper castes, where the upper castes can be up to 90% Western Eurasian, (according to Razib's inference of the Asian reaching a low of 10% in the North West) and some being less Western Eurasian, particularly in the East and South, where the lower castes can be ~60% Western Eurasian?

Like I stated earlier, even the National Geographic Geno 2.0 project substantiates these numbers to some extent, stating that the average NW South Asian is 75% Western Eurasian and 25% SE Asian.

The Western Eurasian value for certain NW Asian groups reaching a high of 90% (according to Razib) would also make sense, considering that ASI reaches a low 14% value in NW groups according to the Eurogenes K7 run.

Are these inferences correct? I'd greatly appreciate some more insight into this matter.

Just wanted to get a different viewpoint based on the 23andme data and Razib's analysis.