How to compute Hudson's/Bhatia's FST in R OR with vcf?

Question

[S]

Hi everyone,

How can I compute hierarchical Fst with Bhatia's/Hudson's estimator using a vcf as input?

My data is structured like this: there are individuals within sampling sites, and sampling sites within groups. My vcfs contain SNP data (~1000 loci for ~250 individuals). I want to compute Fst among sampling sites within groups, AND Fst among groups.

Ideally I could do this in R. I've seen the KRIS package for this, but it can only compare two populations at a time. Multiple authors have mentioned using "custom scripts" to do this but have not published them. I am currently waiting to hear back from one.

I've also looked into using angsd but I'm struggling to understand the workflow for this as it is primarily used for genotype likelihood data. I think I could generate genotype likelihoods with my vcf, and then create an SFS and then the Fst function? But I'm really not sure- haven't used this software before.

Lastly, I think plink might work? But again, I'm struggling to figure out how to write out the command.

Any help would be appreciated! Thanks so much.

Answer 1

[bk11]

Please see in the post below where it had been discussed before-

Error: --fst requires at least two nonempty clusters.

Also you can use the site below for that propose-

http://spsmart.cesga.es/search.php?dataSet=engines

[S]

Thanks for this. I'm trying to make this work and keep getting errors. First, is plink able to calculate fst at multiple hierarchical levels? My data contains individuals within sampling sites within urban/rural groups, and I'd like to understand the genetic differentiation at all levels.

Here is the head of my .txt file with sample IDs, sampling site IDs, and group IDs:

1.2_1A10_sorted MW067   Urban
1.2_1A11_sorted MW079   Urban
1.2_1A1_sorted  MW040   Urban
1.2_1A2_sorted  MW047   Rural
1.2_1A3_sorted  MW002   Urban