tomahawk import does not generate a twk file about tomahawk

[mklarqvist]

Hello @biojiangke,
Can you check and make sure that the variants are diploid, biallelic, and SNVs. These are the only variants kept by tomahawk. Sites with mixed ploidy are also filtered out. Lastly, make sure they are not filtered out because of how you parameterized the missingness threshold.
If you should still have some variants could you send me your VCF please?

[biojiangke]

Hi,

Thanks for the quick response. I tried the following:

Filter out non-biallelic variants:

vcftools --gzvcf locus.vcf.gz --min-alleles 2 --max-alleles 2 --recode --stdout | bgzip -c > locus.biallelic.vcf.gz

Filter out indels:

vcftools --gzvcf locus.biallelic.vcf.gz --remove-indels --recode --recode-INFO-all --out SNPs_only --stdout | bgzip -c > locus.biallelic.snp.vcf.gz

I can confirm the filters as I counted the number of variants before and after filtering and spot checked a few regions. The resulted VCF still does't import for tomahawk, even trying missingness filter from 0.1 to 0.9. Plus, the working VCF have non-biallelic and indels, and tomahawk ignored them and imported the "good" variants. So I don't think those are causing the problem.

I'm bond by regulations and would not be able to share the VCF. But I'll do more tests and report later if I find anything.

[mklarqvist]

This should no longer be happening as of release 0.7.0 as we now rely on htslib to do the parsing.