Referee 1's review (John Hawks):

Review of Li et al., "Ethnic related selection for an ADH variant within East Asia."

Recommendation: Minor revision

The introduction should clearly differentiate this paper from the previous study, reference 12, which seems to overlap substantially in content, although the current paper has more samples.

There seems to be some confusion in the introduction concerning the power of tests of selection. For example, the text says,

"in some Austronesian (AU) populations, the derived allele frequency of the functional polymorphism is very high, but there is no evidence showing that the functional allele ADH1B*47His underwent selection[12]. Thus, not only is the selective force unknown, but selection cannot explain the frequency of ADH1B*47His in all populations in which it is high."

These observations are of course directly related to each other: the power of the test of neutrality depends on the derived allele frequency. Since the cited study found very high values of EHH for the Austronesian populations, it is not clear why this manuscript thinks that supports the idea of no selection.

Later in the text, the authors refer to "weak selection," when surely they mean "weak evidence of selection."

It seems to me that the important paragraph is the first paragraph under "The selective region and allele," which considers that the alcoholism-related variant is not the target of selection, without linkage to the promoter variant. This deserves more attention: is it possible that the promoter was selected irrespective of the ADH1B*47His allele?

This also raises a question about the relevance of the linguistic comparisons -- if the selection is on the promoter allele, and evidence of selection is positively associated with promoter allele frequency, then what difference does it make that the evidence of selection appears less well correlated with ADH1B*47His frequency among language groups? The confounding variable looks like linkage, not culture.

I agree that the cultural history must have been important, but without a better idea of the functional relation between these variants and fitness it will be hard to specify. And it would require the linkage between markers in the core haplotypes to be resolved anyway.

This statement:

"The population history mode in our simulation was designed according to the
migration history, archaeological discoveries and the environmental changes in East Asia"

is entirely ambiguous. What did this entail?

I think the paper would benefit from some clear explanation of the issue of alcoholism and linkage to these alleles in relation to the cultural variability discussed. I would expect an ordinary reader to think, "Hmm...I guess they're saying these cultures differ in the fitness cost of alcoholism." Maybe they do, but I don't think that's what the manuscript is getting at.

**********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication the manuscript has been revised in light of these comments and to address other editorial requirements.