September 14, 2008

Y chromosomes of Bayash Romani

Once again, the 0.00069/locus/generation rate is used in this paper, and hence its estimated ages are wrong. The given Y-STR variance for haplogroup H1a in Table 2 is 0.06, which corresponds to an age of ~800 years.

It's interesting though, that Zhivotovsky is a co-author of this paper which states that:
A recent refinement of E1b1b1a-M78 by novel biallelic markers indicates that its subhaplogroup E1b1b1a2-V13 is the most common in Europe (Cruciani et al., 2007). In fact, E1b1b1a2-V13 originated in Western Asia about 11 KYA and expanded in Southeastern Europe about 4.5 KYA, not in connection with the spread of agriculture as traditionally assumed, but rather at the beginning of the Balkan Bronze age, as a consequence of the in situ population increase in the already populated territory (Cruciani et al., 2007).
and he was a co-author of King et al. (2008) which stated that:
The calculated expansion time of haplogroup E3b1a2-V13 in mainland Greece is 8,600 y BP at Nea Nikomedeia and 9,200 y BP at Lerna/Franchthi Cave and is consistent with the late Mesolithic/initial Neolithic horizon. These dates exceed those reported previously for Europe (Cruciani et al., 2007) that date to the Bronze Age. This discrepancy arises mainly because of differences in the choice of mutation rate used.
Peter Underhill was also a co-author of the latter study, and also of the recent paper on Sicily which used germline mutation rates and:
The estimate of Time to Most Recent Common Ancestor is about 2380 years before present, which broadly agrees with the archaeological traces of the Greek classic era.
Mesolithic - Early Bronze Age - classical Greek. Three completely different ages using three different mutation rates: a mutation rate 3.6x slower than the germline rate => Mesolithic. A mutation rate 2.4 to 2.8x slower => Early Bronze Age. A germline mutation rate => classical Greek.

My most recent take. I'll be much surprised if E-V13 turns out to be anything other than 2nd millennium BC in the Balkans.

American Journal of Physical Anthropology doi: 10.1002/ajpa.20933

Dissecting the molecular architecture and origin of Bayash Romani patrilineages: Genetic influences from South-Asia and the Balkans

Irena Martinovi Klari et al.


The Bayash are a branch of Romanian speaking Roma living dispersedly in Central, Eastern, and Southeastern Europe. To better understand the molecular architecture and origin of the Croatian Bayash paternal gene pool, 151 Bayash Y chromosomes were analyzed for 16 SNPs and 17 STRs and compared with European Romani and non-Romani majority populations from Europe, Turkey, and South Asia. Two main layers of Bayash paternal gene pool were identified: ancestral (Indian) and recent (European). The reduced diversity and expansion signals of H1a patrilineages imply descent from closely related paternal ancestors who could have settled in the Indian subcontinent, possibly as early as between the eighth and tenth centuries AD. The recent layer of the Bayash paternal pool is dominated by a specific subset of E1b1b1a lineages that are not found in the Balkan majority populations. At least two private mutational events occurred in the Bayash during their migrations from the southern Balkans toward Romania. Additional admixture, evident in the low frequencies of typical European haplogroups, J2, R1a, I1, R1b1b2, G, and I2a, took place primarily during the early Bayash settlement in the Balkans and the Romani bondage in Romania. Our results indicate two phenomena in the Bayash and analyzed Roma: a significant preservation of ancestral H1a haplotypes as a result of considerable, but variable level of endogamy and isolation and differential distribution of less frequent, but typical European lineages due to different patterns of the early demographic history in Europe marked by differential admixture and genetic drift.



McG said...

There has been a thread on DNA forum started by Derek that has been discussing the, at least two, different mutation rates: germline and ZUL. What it has boiled down to is this point. Why should autosomal rates be different than STR non-recombing rates for the same types of molecules (tri and tetra nucleotides)?? In the 04 paper by ZUL, they also showed in addition to population ages, that the .00069 evolutionary rate is the same for autosomal microsatellites

Maju said...

Once again, the 0.00069/locus/generation rate is used in this paper, and hence its estimated ages are wrong.

Stubborn! :)

John said...

Once again, the 0.00069/locus/generation rate is used in this paper, and hence its estimated ages are wrong.

True. The rates fit some of the time, but they are well off the mark in other cases.

Andrew said...

If I read between the lines correctly, this Roma group show a form of E-M78 which is not E-V13? Could they be M123+ which is known to exist in the Balkans? Could the Roma be the vector for the entrance of non V13 into the Balkans? I have not seen the article, so this is just a speculative question.

Dienekes said...

If I read between the lines correctly, this Roma group show a form of E-M78 which is not E-V13?

No, it is E-V13.

McG said...

John said the rates are well off the mark? I don't know. On the DNA forums thread there was a proposal to test which "rate" was right by estimating the time the Amerind came to N.A. Well, how do you do that?? TMRCA analysis of a group of entries tells you nothing about where the group was at physically at the time of the TMRCA. You do usually colalesce to the oldest entry, but what does that tell you? Misinterpretation of what TMRCA tells you may be more of the problem than right or wrong rates??? My point on the Amerind is that you could well coalesce on a time prior to the entry to NA, if the diversity of the entries was sufficient??

terryt said...

And I would expect the entries to be reasonably diverse.