FST: allele-counts or allele-frequencies?

Question

[Adrian Pelin]

Hello,

I am trying to calculate genome-wide FST pairwise values for every gene between all my isolates which I sequenced using NGS. I am offered the default method using allele-frequencies coming from Hartl and Clark (2007), and the allele-counts method by Karlsson et al. (2007).

Would anyone be able to tell me what are pro's and con's from experience of one method versus the other?

Adrian

Answer 1

[Zev.Kronenberg]

If you have genotypes, rather than pooled data, I would use allele-count methods. Allele frequency differences do not utilize heterozygosity information, which can provide additional information.

Weir and Cockerham 1984 is a popular method. I have implemented it in wcFst. It takes a VCF as input.

https://github.com/jewmanchue/vcflib/wiki/Association-testing-with-GPAT

[Adrian Pelin]

I have sequenced populations of spores, so each isolate is a pool of all individuals in one population, and there is no way to sequence individuals. After analyzing the data it seems that the spores are propagating clonally, so most of the variation can be explained by heterozygosity. Which method do you think would be best to use?

Also, for your software, what do you recommend as a variant caller to obtain VCF? I currently use FreeBayes