I’ve mentioned a few times that the Reich lab has been finding suggestive evidence for admixture between indigenous South Asians and a West Eurasian group on the order of ~3,000 years before the present. The modal explanation is probably an Indo-Aryan intrusion. Dienekes used rolloff in ADMIXTOOLS to repeat these general findings. Specifically, he found signal for an admixture event analogous to one between non-Brahmin South Indians and Northern Europeans. I say analogous because I do not mean to imply that the admixture was exactly of this form. Rather, there are general resemblances in the genetic profiles across the four groups (i.e., Orcadian & North Kannadi, and population X and Y which merged to form South Indian Brahmins).
The only thing I would add is that the North Kannadi, representative of non-Brahmin South Indians, are themselves a mixture between West Eurasian and an indigenous South Asian population. And, looking at the patterns of admixture in Harappa, it seems to me that South Indian Brahmins themselves have some admixture which postdates their arrival to South India, which is probably 1 to 2,000 years before the present. In the specific case of South Indian Brahmins my model would be:
1) West Asian + Ancestral South Indian = Population A
2) Population A + Indo-Aryan = Population B
3) Population B + Population A [South Indian] = South Indian Brahmin
I would weight the first admixture event heaviest, and the two second ones of approximately equal strength. Though the third is a secondary admixture, and so I wonder if the signal would be obscure.