Estimating Nucleotide diversity (Pi) from VCF file in polyploid organism (tetraploid)

Question

[Michael]

I am trying to estimate some basic population genetics parameters such as Pi, Tajima's D, Fst, etc. from a VCF similar to what is described here. However, the organism is polyploid, so vcftools won't do the job (Error: Polyploidy found, and not supported by vcftools: NC_001133.9:24). I am assuming there is another quick solution for this but am unable to find it. Hope you can help me with this.

Thought about converting the variant calls to diploid, ploidy shouldn't really matter, should it? Or re-running variant calling with ploidy set to 2. Variant call pipeline used is GATK HaploType caller -> Joint haplotypes.

[dthorbur]

The readVCF function from the R package PopGenome says it supports polyploid individuals. It can calculate all the metrics you list and more.

It also seems you could change ploidy with little negative effect in GATK, so long as you aren't calling variants in the germline. GATK released an FAQ about it recently.