(2015-May-05, 22:19:30)Chuck Wrote: I'm stuck on this question, though -- which I can't seem to get an answer to:

............................................................................

Dear Blah blah blah

I am writing to see if either of you could help with a population genetic question, a clear answer to which has been eluding me.

For context, I am writing a paper on the biological race concept in relations to our species.

I would like to determine, based on genetic differentiation values (e.g., SNP Fst values), what the expected between-group-variance would be for an average quantitative trait owing to neutral divergence.

Would it be:

(a) ~Fst (treating this as an F-ratio that could be converted into mean standardized differences)

(b) ~2*Fst ("")

© Something else.

Much of the literature discusses the matter in terms of Qst, defining this as GB/(GB + 2GW). And Qst is said to come out to Fst.

But it's not clear to me if, for diploids (where roughly half of the genetic variance is trapped within individual), the expected quantitative trait F-ratio would be equal to 1* or to 2* Fst.

I stumbled across this discussion by Whitlock (2008) ( "Evolutionary inference from QST")

"Does FST = QST for neutral traits?

The calculation of QST for a trait requires two quantities: the additive genetic variance of the trait within a population (VA,within) and the genetic variance among populations (VG,among). For diploids, QST is calculated as

Qst = GB/(GB + 2GW)

For haploids, the same equation applies, but without the '2' in the denominator. [That '2' for the diploid case comes from the fact that the quantitative genetic variance among populations is proportional to two times FST (Wright 1951).] "

...

It seems to say that the expected between population quant variance would be 2*Fst.

This strikes me as being rather high, though.

To take human continental natural divisions as an example, the Fst SNP value is ~ 0.12.

This would then given F-ratio = 2*0.12 = 0.24 which would be equivalent to an average d-value > 1.00.

From what I recall the typical phenotypic F-ratio, though, is around 0.10 (i.e., in dental and craniometric traits).

Blah blah blah...

.............................

You wouldn't have any idea, would you? I asked Henry Harpending who is more familiar with the Qst literature (inferences of neutral phenotypic divergence from genetic data), but he wasn't sure. I got the basic idea. But I wasn't able to be precisely clear in my discussion because I couldn't resolve this issue -- so I had to talk around it.

(Ya, I understand your apples to oranges point -- I would still like to know what the predicted quantitative trait variation would be due to neutral divergence. See this paper for context: Leinonen, et al. (2013). QST-FST comparisons: evolutionary and ecological insights from genomic heterogeneity.)

First of all, the set of 315 Fst values that I calculated using VCFtools (which employs Weir and Cockeram Fst formula) on 1000 Genomes phase 3 data for 26 populations can be seen here (https://docs.google.com/spreadsheets/d/1...sp=sharing ). I report Fst for 1st and 21st chromosomes (columns C and D). They are practically identical (r=0.995) so either can be used to represent the whole genome. Note that these include SNPs and indels. If you use these Fst values in your paper, please cite my last article (http://dx.doi.org/10.6084/m9.figshare.1393160 ) because they are in the supplementary material there.

THERE IS INDEED MUCH CONFUSION ON INTERPRETING FST AS RELATIVE BETWEEN POPULATION VARIANCE.

It appears that the expected BETWEEN population variance should be 2*Fst, after correcting for the inbreeding coefficient.

Sarich and Miele (2004) write "Lewontin had noted that 85% of the genetic variability was among individuals within populations, and only an additional 15% was added when individuals in different populations were compared [...]. The point is that we are diploid organisms, getting one set of chromosomes from one parent and a second from the other. To the extent that your mother and father are not especially closely related, then, those two sets of chromosomes will come close to being a random sample of the chromosomes in your population. And the sets present in some randomly chosen member of yours will also be about as different from your two sets as they are from one another. So how much of the variability will be distributed where?

First is the 15% that is interpopulational. The other 85% will then split half and half (42.5%) between the intra- and interindividual within-population comparisons. The increase in variability in between-population comparisons is thus 15% against the 42.5% that is between individual within-population. Thus, 15/42.5=32.5%... "

So your hunch was right in telling you that Fst underestimates the between-population comparisons by about a factor of 2 (in reality it's a little less than 2 if we account for inbreeding).

Sarich, V. and Miele, F. (2004). Race. The reality of human differences. pp 168-169.