I-L621/Dinaric I2a origins: from the East or the West?

Before my recent comment about a sample from Kılınç et al. Science Advances (2021) fades into oblivion:

Cis-Baikal Medieval irk032 (1277-1393 calCE), from Zarubino site, burial 3 (Angara river, Irkutsk Oblast), of the Cis-Baikal_673_557_BP cluster, is of mtDNA D6c, and most likely of hg. I2>CTS2257>L460>P37>M423>S2770>CTS5375>CTS7213>L621>CTS10936>S19848>CTS4002>CTS10228>S20602>Y4460>Y3106>A6106, without any controversial calls.

(A) PCA calculated using a set of world populations. Ancient individuals were projected onto the inferred PC space (see table S2 for information about individuals). Arrows indicate the direction of population changes in time. Asterisk denotes being published in (16). (B) A subset of ADMIXTURE result for K = 14 clusters showing the ancestral composition of investigated ancient individuals. Modified to remark the position of Irk032. See more information here.

This is weird, since no Slavic expansions are expected there by that time. Its position in the PCA is also weird, but the genome seems of enough quality. The radiocarbon date of this sample from Zarubino in the Angara Valley is published in Shergin & Ivanov (2019).

A surprising finding, given the previous report of an Early West Slav from Poland and Serbia (almost coeval with Early Hungarians), and especially of the leaked sample from BA Bulgaria, of which I had my doubts. If the latter were a (pre)I-L621 sample, then some back and forth through the (forest-)steppes would be conceivable.

Modern descendants suggest a star-like phylogeny starting in Central-West Europe and continuing (through the Lusatian culture?) in SE Europe (with Proto-Slavs?), but it is evident from the data on other haplogroups that the early formation dates and TMRCA, huge overwhelming differences in the number of testers (more from Western Europe, less from Eastern Europe, and very few from Asia), as well as multiple migration waves from the east, might have distorted the original picture of steppe-related migrations.

SNP Tracker’s migration path for I2-L621, inferred from modern descendants tested in FTDNA

This East Siberian Early Medieval sample, together with the Hungarian Conquerors of hg. I2-L621(CTS10228>S20602), including one leader (none with genome-wide analyses), suggest – yet again – that (part of?) these lineages were also related to (forest-)steppe nomadic expansions, regardless of the ethnolinguistic groups and traceable geographic origin or TMRCA of the majority of modern descendants.

The only good reason for this haplogroup to appear that far east despite an origin in the west would be the same international connections that allowed for the appearance among Mongols of hg. R1a-Z92 (DA29, from Karasuyr, Kazakhstan), or of R1b-U106 (TAV007, from Tavan Tolgoi, Mongolia), during the same approximate period, although the former is clearly an outlier among the Golden Horde, and the latter is not confirmed with BAM analysis yet.

Until more samples get officially published, it’s impossible to say where the majority of European I2-L621 subclades (i.e. I-Y3120/S20602) came from ultimately. In fact, without some I2-pre-L621 sample, it might impossible to say with certainty.

So back to square one, where some I2-L621 (like some R1a-Z280 subclades) seem to be linked to recent steppe-related (Iranian, Uralic, “Altaic”) related migrations. In chronological order, some posts regarding the publication of I2-L621 samples:

Magyar tribes brought R1a-Z645, I2a-L621, and N1a-L392(xB197) lineages to the Carpathian Basin

“Dinaric I2a” and the expansion of Common Slavs from East-Central Europe

Vikings, Vikings, Vikings! Hordes of high quality ancient DNA