VCF incorrectly specifies phased genotypes

[davidlougheed]

Hello, I ran LongTR on some data, specifying a .vcf.gz file as output, but ran into the following issue with said output:

• the contents of the file are a plain-text VCF, not a bgzipped VCF, despite it ending with the .vcf.gz suffix
• the genotypes are separated by a |, not /, which indicates phased genotypes, but I did not provide a SNP VCF for it to phase with (the BAM is technically phased though, so maybe it's automatically using those haplotags? but I'm pretty sure it isn't.)

[heliziii]

Hi David,

Thank you for using LongTR!

So the output name is ending with vcf.gz but it is not bgzipped? If you run bgzip -d on it, does it give you an error?

The genotypes are always separated by |, regardless of phasing information. It is on our list to correct this, I apologize for the confusion.

Best, Helia

[davidlougheed]

Ah my apologies, I got confused; my editor was automatically decompressing them on the fly - I'm very sorry. I'll re-title the issue to just refer to the phasing for future reference for others.

[heliziii]

No worries!