R-M12205

The Balt / Iran split in language also has a genetic component, which is being hashed out among the mostly-Semitic Tribe. Specifically there is an old R1a signature in Levites, who then assuredly are not Levides save by adoption.

Jewish Genomics has the scoop. Which it is, because @FamilyTreeDNA is supposed to get more-formal publication later.

A "Tajik" from Pamir shares R-Y2619 ancestry but was clearly no Jew. The relevant Levites are now reclassified: R-M12205, counting down from AD 750.

So far the Levites in question are Ashkenaz, which is why - beforehand - it was a tossup if R-Y2619 were just East European. Pamir is now Badakhshan so decidedly Iranic. Something like Avestan, Parthian, Bactrian et al.. 'Twould be later that the region became Persianate (or, before that... Buddhist).

"Dienekes" Lazaridis thinks some north-Iran tribe off to itself from 700 BC to injection into Judaea, whenever that was. They'd have drifted west. Lazaridis could call them Medes.

If Lazaridis be right that the injection was so early: I find personally difficult to understand how this bloodline got restricted to the Ashkenaz, "the Scythia" to Greeks and "Saka" to Seistanis (and it's probably not them anyway).

Other commenters are reviving the Khazar hypothesis - was Elhaik right for the wrong reasons? Or maybe this is old Radhan, spreading out to the Caspian shores to the Tatarstan around Kazan. I mean, I know only my femme ancestry and that is decidedly Oriental.

I suspect that we have an incomplete record of male Y-DNA from Sephardic and Mizrahi populations. To be kept in mind: occasional "Ashkenaz" signatures can be imported, like in AD ~1666.