An Estimate of Steppe Ancestry In South Asia

Since this question always comes up at some point, I decided to do a rough back-of-the-envelope calculation of the % steppe across the Indian subcontinent. The way I did it was by taking Pakistan, Bangladesh, and India, and estimating the average percentage from the caste breakdowns (e.g., UP is 20% “upper caste” and 20% “Dalit” and 60% neither, with fractions of steppe/Sintashta about 30%, 10%, and 15%, respectively).

So the final number I came back is that 14% of the ancestry in modern-day South Asia is from the steppe in the form of people descended from Sintashta pastoralists. . . . . You can judge whether that’s significant or not. Additionally, it looks like closer to 20-25% of the Y chromosomes are derived from these people.

From Razib Khan at Brown Pundits. 

He gives himself a margin of error due to methodology issues of about ± 2 percentage points.

In general, Brahmins have more steppe ancestry, while this percentage drops (sometimes particularly sharply outside of NW India) with lower caste. It is more common in the Northwest and reduces in frequency (in non-Brahmins, at least) in a more or less clinal pattern from there. 

There are no examples of contemporary people native to South Asia who lack steppe ancestry entirely (although people without steppe ancestry are found in ancient DNA, for example, Harappan ancient DNA from ca. 2200 BCE). 

The data imply a source for steppe ancestry that was about 80% male, and certainly disproportionately male.

The fact that so much geographic and caste structure in steppe ancestry percentages exists at all is remarkable for a source of admixture with so much time depth (and which, in particular, significantly precedes the hardening of endogamy boundaries for jati in India).

The comments explore some finer details. 

About 35% of autosomal DNA in South Asia is autochthonous (i.e. native to India, at least in the Holocene era of the last 10,000 years or so) as is a large share of South Asia's Y-DNA and mtDNA. The lion's share of the rest is West Eurasian, mostly derived from ancient Iranian farmers, except in Northeast India and in Bangladesh where there is significant East Asian and Southeast Asian ancestry (mostly in people who speak Tibeto-Burman or Munda languages, or in people who are geographically adjacent to these peoples).