How should I convert VCF to Phylip format?

[silvewheat]

Hello,

Since HyDe use sequence data in Phylip format as input. But I only have a VCF file. I'm not sure what is the best way to convert a VCF to sequence data. Specifically, I have the following questions:

1. Should I keep the non-variant sites (that not exist in VCF file but in the reference fasta file) and genrate a consensus sequence? Or I just need use only sites in the VCF file?
2. Should I convert each diploid individual to two haploid sequence? Or I can just treat diploid as one individual and use the ambiguous nucleotides in heterozygous sites?

Beast, Yudong

[pblischak]

Hi Yudong,

For converting to Phylip format you should be able to just use the variant sites in your VCF file. Also, I think that keeping both alleles for the diploids would be better as long as you aren't trying to run the individual-level tests. Because HyDe assumes independence among sites, it shouldn't matter which allele you assign to the two haploid individuals you make from the diploid VCF file. If you are wanting to run the test on individuals though, then I think using ambiguity codes for heterozygous sites and coding each individual as a single consensus sequence would be best

Hopefully this helps but let me know if you have any other questions

Paul

[silvewheat]

Hi Paul,

Thanks for you answers. I'll try as your suggestion.

Best, Yudong

[zsdxgl]

There are always missing genotypes in VCF files. How to deal with these sites? Whether the program HyDe can handle "N"?

[pblischak]

Hi @zsdxgl -- HyDe can handle missing data ('N'). It does this by trying to integrate over the uncertainty that missing data creates by assigning 0.25 to each of the possible nucleotides