Error while subsetting VCF – error doesn’t check out with (z)grep

Error while subsetting VCF – error doesn’t check out with (z)grep


I’m using bcftools view -s to subset a VCF.gz file. I ran into an error:

[E::vcf_parse_format] Number of columns at chr9:44897051 does not match the number of samples (90 vs 99)

To look at this site, I ran a grep and picked the line. Then, I checked out the number of columns in the line and it turned out to be 108 (99 samples + 9 fixed columns).

What could be going on here?

I’m running a grep right now to pick the lines around this entry to see if they look odd, but barring that, could something else be wrong? I am worried because I also see this warning

[W::bgzf_read_block] EOF marker is absent. The input is probably truncated

which is sort of baseless (I can’t figure out why the warning even exists, given the script that generated the file ran to successful completion – see here: EOF marker absent in VCF – can this be safely ignored?)

I’d appreciate your inputs on this situation.


I looked at 3 lines around the chr9:44897051 entry and all of them have 108 columns. I don’t understand what could be happening.




Source link