The above figure shows the frequencies of Y chromosomal haplogroups of men of South Asian who claim to be descended from the prophet or his tribe, as cross-referend with their surnames. The “Non-IHL” category indicates those who are not of these honored lineages.
The paper from which I drew the data, Y chromosomes of self-identified Syeds from the Indian subcontinent show evidence of elevated Arab ancestry but not of a recent common patrilineal origin, actually somewhat support the idea that these people descend from Muhammad or the Quraysh or the Ansar.
I think this is wrong.
But first, why do think these data results show Arab affinity? The “IHL” lineages have a higher proportion of haplogroup J, the most common haplogroup among Arabs. J is not exactly rare in South Asia (lots of <<<Brahmins>>> who are not sons of Indra have it because they are the scions of cunning Dasa priests), but there’s clearly a frequency discrepancy.
And yet this paper was published in 2010. We now know through various tests of confirmed descendants of Muhammad, and who descend in the male line from his cousin Ali, that they carry a branch of haplogroup J1.
Even among the Syeds, most do not descend from Muhammad assuredly. There are nearly as many scions of Lord Indra, R1a1, as those who bear haplogroup J. Of the J’s within the Syed community, I think the most likely scenario if they are not South Asia is that they are Iranian. J is found at frequencies of 35% in Iran, and Iranians, along with Turks, were the most common migrants into South Asia.
In other words, the Syeds of the Indian subcontinent are the sons of magians, not Muhammad.


There are ~1.5 billion of us and only a few tens of millions of <<<them>>>

