iqbal-lab-org / cluster_vcf_records Goto Github PK

This will make creating bioconda recipes for packages that rely on this easier/possible

I've found a bug from running the module on our Pf3k vcf. The issue is: the output from clustering the records in that vcf is currently a vcf which still has overlapping records.

I've traced the putative cause down and am implementing a new unit test which covers it.

Will update on progress

Currently the module loses the headers in the vcf.

For gramtools this is a problem if you make a clustered vcf, convert it into a prg, and genotype several samples again the prg. If you then want to combine the vcf's, you can run into problems, eg bcftools needs the vcf to be properly formatted.

Another thing is maybe to check that the vcf is properly formatted even without losing the headers.

I have been struggling with a 'KMER=' INFO column presumably from cortex, undocumented in the headers.