Missing lots of SNPs when run CONVERTF in EIGENSTRAT 5

Question

[fu_entomology]

Hi everyone in the community,

I am trying to analyze SNP data for an insect population genetics study. We did a RAD-tag sequencing with around 300 individual insects, there is no reference genome. We aligned shorts reads among samples and called SNPs in a VCF file.

I converted the VCF file to .ped and .map formats using vcftools, and loaded them to EIGENSTRAT 5, I have had success with EIGENSTRAT 5 before with unplaced contig data, as you know the chromosome number might get messed up and not recognized by some programs.

However this time, after I ran the CONVERTF, only 100 SNPs showed up in the .eigenstratgeno and .snp file. My original plink files have 6700 SNPs. I ran again with subset of the data with fewer SNPs and fewer sample number, the problem remained.

Wonder did anyone here encounter the same issue and how did you fix it? Lots of thanks!

Answer 1

[Philipp Bayer]

What is in the phenotype column (column 6) in the ped file? If it's -9 CONVERTF will remove it, try setting that column to 1

Edit: Oh, 17 months ago created but updated today, let's hope someone else may find this answer useful.