calculating allele frequency difference or Fst value ?

Question

[seta]

Hi all,

I would like to determine allele frequency difference of a set of SNPs (say 2500 SNPs) between various populations as my first experience in this regard. I think of using the statistical test (Fisher’exact or chi-square test) with the allele count for it. However, I see the estimating Fst value mostly used. Could you please let me know which method is more appropriate for finding varying SNPs among the populations and why?

Also, please kindly tell me what is the suitable tool for calculating Fst, there lots of tools, which one do you suggest? and what’s your advice for finding statistically significant outlier Fst?

Thanks in advance

Answer 1

[jfo]

Hello Seta,

Here are some quick answers to your questions.

You can always try both methods to determine genetic differences in your populations of interest. Which method is more appropriate will depend on the hypotheses you want to test. For estimating genetic differentiation between populations, I think it is important that you report FST estimates.

For estimating FST, Weir and Cockerham (1984) is one of the more commonly used estimators. This can be computed using genepop or various R libraries like genetics or diveRsity for better handling of the input and output (because programming).

Outlier test can be usually done using BayeScan, Lositan, and Arlequin. Some use them individually, but some also run them all and report the potentially outlier SNPs common to all approaches. How you determine the outlier SNPs will depend on how much stringency you want to apply to your data set.

Good luck to your analysis! :)

Answer 2

[paolo002]

Hi Seta Personally I would recommend ANGSD (http://www.popgen.dk/angsd/index.php/ANGSD). It can calculate Fst in a straightforward way.