R1b1b2 Cohanim

A new paper by Hammer et al. (Extended Y chromosome haplotypes resolve multiple and unique lineages of the Jewish priesthood) focuses on the J-P58* lineage among Jewish priests (Cohanim). But the paper adds some data to the understanding of R1b1b2 also.

Among the 215 Cohanim haplotypes published in the paper, 13 are R1b1b2 (one is mis-identified as R-P25). The paper doesn't treat these R1b1b2 in a detailed manner, and I think reaches a suspect conclusion:

In contrast, our
network and divergence time analyses suggest that R-M269
chromosomes entered the Cohanim population via several
“migration” events, and do not represent a single Cohanim
founding lineage. For example, divergence time estimates
are much older for Cohanim R-M269 chromosomes (>10
kyears) than for the three Cohanim lineages in haplogroup J
discussed above, and median-joining networks of Cohanim
R-M269 chromosomes lack a modal haplotype and show
many unrelated singleton haplotypes that are interspersed
among Cohanim and non-Jewish samples (Figure S4).

However, a closer look at the R1b1b2 data from the paper is revealing

First, take a look at a table of pairwise genetic distances for the 13 haplotypes. I color-coded the GD values, so the red ones are the most closely related. You can quite easily see that there are two clusters of R1b1b2 Cohanim, with the first cluster having six haplotypes and the second cluster having three.

The first cluster is older (with a TMRCA of around 1600 years ago, or about 400 A.D.) By most accounts of Jewish history, the MRCA for this cluster was either contemporaneous with or slightly pre-dated the largest diaspora movements. The second cluster appears younger (maybe 1200 AD). Both of these clusters have members in the Jewish R1b project at FTDNA, so more detailed analysis with longer haplotypes should be possible. Both clusters are likely R1b1b2a* (aka R-L23).

Another phenomena comes from a graphic representation of the GD between pairs of haplotypes. For this comparison I show three series.

Look at the red series first. This represents 100 men chosen at random from the general R1b project at FTDNA. It shows a relatively smooth distribution, peaking at a GD of 8. In other words, the modal distance between two random haplotypes using the same 22 markers as the Hammer paper is 8.

Next, look at the blue series. These are the 13 R1b1b2 Cohanim from the new paper. Their distribution has some significant differences. One is a substantial excess of closely related haplotypes. This means that the R1b1b2 Cohanim are not merely or entirely the result of some random admixture, but rather that a significant fraction of the Cohanim R1b1b2 lineages have pre-diaspora or diaspora-era connections.

A second difference is the second, larger peak at a GD of 11. This is reflecting an early seperation of R1b1b2 lineages around 4,400 years ago, most probably in the Near East. You can also see this in the green series, which shows the GD distribution of all the Cohanim/non-Cohanim pairs (1300 pairs, in fact). The Cohanim are more distant from random R1b1b2 European men (on average) than R1b1b2 European men are from each other. This reflects the fact that 80% of the Cohanim men appear to be L23+ L51- while most European men (the 100 "random" ones mentioned earlier) are not only L51+ but also L11+ and either P312+ or U106+.

R1b1 Neighbor-Net

I have used SplitsTree4, an excellent piece of phylogenetic software that I regret not learning earlier, to produce a phylogenetic network using the Neighbor-Net method.

I'll try not to bore you with the nitty-gritty, but Neighbor-Net is a type of distance based phylogenetic network. I think of it a as a hybrid between the median-joining networks produced by Fluxus and the Neighbor-Joining trees produced by programs like Phylip. I encourage you, if you don't mind the tecnical details, to read the paper by Bryant and Moulton presenting Neighbor-Net.

Unlike a phylogenetic tree, which shows the most likely or most parsimonious tree, a network shows many (or all) feasible trees in one 2D representation. In effect, it allows us visually represent the ambiguity we have about which of the feasible trees is the the true tree.

In this case, the input was pairwise distance matrix constructed (using Dean McGee's Y-Utility) from STR haplotypes at Ysearch and the R1b1b Haplogroup Project. It shows the relationship of a representative sample of the known sublcades of R1b1 (R1b1b in blue and R1b1c in red) to the other clusters of R1b1* for which no SNPs downstream of P25 have yet been identified.

Each of sublcade and/or cluster can be seen to "radiate" out from the center of the network. Each node is labeled with a ysearch ID (except the R1b1b folks, which are labeled with FTDNA kit numbers). I don't want to overstate any conclusions, but generally speaking the longer the "neck" is for each cluster the more time passed between its split from the group and the era in which the MRCA for that subclade/cluster lived.

Why is this blog?

This blog will be a place for me to write about Y-DNA haplogroup R1b.

I will write about undoubtedly write about the super-popular topic of subclade R1b1c but will probably focus more energy on other subclades, like R1b1* and R1b1b.