Perl script does not include VCF entry unless PASS is in specific column

iqbal-lab-org / gramtools

Genome inference from a population reference graph

MIT License

92 stars 15 forks source link

Perl script does not include VCF entry unless PASS is in specific column #102

Closed ffranr closed 6 years ago

iqbal-lab commented 6 years ago

I think this is correct behaviour btw - a VCF line needs PASS in the filter column to be valid. People stuff all kinds of shit into VCFs and mark the bad lines in that column

ffranr commented 6 years ago

Is the following behavior correct?

VCF filter column identified regardless of position.
If filter column absent, include all entries.
If filter column present, include entries which have PASS as filter column values.

I will modify the title.

iqbal-lab commented 6 years ago

Do you mean column absent or for a given row, if there is no entry?

iqbal-lab commented 6 years ago

Column is mandatory, my VCF (that triggered this) was malformed

ffranr commented 6 years ago

I didn't realize that the filter column was mandatory. I've just consulted the VCF (v4.3) specs. I now see that the filter column is mandatory and that the columns ordering is fixed.

The perl script is correct and that VCF file was malformed. I'll close this issue.