A. Zoossmann-Diskin has a new article in Homo in which he casts significant doubt on the practices of population geneticists publishing on the subject of Jewish origins, with an emphasis on the purported "Ashkenazi Levite" haplotype. While I have also criticized the identification of the so-called Cohen Modal Haplotype whose alleged correspondence with the Aaronic line of priests is invalid, this new article casts some even more serious doubt on a long string of research results. I will comment on the subject after a careful reading of the paper.
The point of the article is that the frequency of the LMH in the original publication is much less than that in the later publication. In the initial publication it is 21% among Ashkenazi Levites and it becomes 42% in the later publication.
1. Significant typing errors occurred in the first publication, and the 21% frequency is widely off-the-mark. Did they publish an erratum, a letter to the editor, an announcement on their website, correcting this error? Or did they allow their result to stand in the scientific literature, even though they knew it had serious errors?
2. They used a biased convenience sample in their second article which just so happened to have twice the frequency of the LMH?
So, either the researchers botched the first publication and kept quiet about their mistake, or they massaged their second sample to increase the frequency of the LMH.
I would be interested in any alternative interpretations.
HOMO - Journal of Comparative Human Biology
(Article in Press)
Ashkenazi levites’ “Y modal haplotype” (LMH) – an artificially created phenomenon?
The article on the Y chromosomes of Ashkenazi Levites (Behar et al., 2003. Am. J. Hum. Genet. 73, 768–779) is the fourth in a series on the Y chromosomes of the three Jewish male castes: Cohanim (priests), Levites (priests’ helpers) and Israelites (lay people). It became apparent that there is a problem with omission of samples when the second article “Origins of Old Testament priests” (Thomas et al., 1998. Nature 394, 138–140) was published. In the fourth article a remarkable 55% of the Ashkenazi Levite samples from the earlier 1998 study are not included. This causes the “Levite modal haplotype” to double its frequency from 21% of the Ashkenazi Levite sample in 1998 to 42% of the Ashkenazi Levite sample in 2003. The authors offer three main explanations:
(1) The studies are independent using different sample sets.
(2) Typing errors and poor quality exclude samples from future studies.
(3) Correction of typing errors means that some samples are classified under different haplotypes.
The explanations offered to the problem of omitting samples from subsequent studies after their haplotypes or partial haplotypes are known, are not convincing. Consequently their sample sets cannot be considered random and non-biased. At the least, these laboratories have bad practices of sample handling and many typing errors, which are enough to invalidate their studies.