VCFtools filter - genotype quality score changes

Question

[AW]

Hi,

I ran the following filtering command

vcftools --vcf test.vcf --out test.test.vcf --remove-indels --remove-filtered-all --max-alleles 2 --max-alleles 2 --recode --recode-INFO-all

However, when I compare the new test.test.vcf compared to the original test.vcf the GQ scores of some snps have changed.

eg

test.vcf

scaffold11  389 .   T   C   .   PASS    ADP=22;WT=0;HET=0;HOM=1;NC=0    GT:GQ:SDP:DP:RD:AD:FREQ:PVAL:RBQ:ABQ:RDF:RDR:ADF:ADR    1/1:123:22:22:0:22:100%:4.7526E-13:0:41:0:0:17:5

test.test.vcf

scaffold11  389 .   T   C   .   PASS    ADP=22;WT=0;HET=0;HOM=1;NC=0    GT:GQ:SDP:DP:RD:AD:FREQ:PVAL:RBQ:ABQ:RDF:RDR:ADF:ADR    1/1:99:22:22:0:22:100%:4.7526E-13:0:41:0:0:17:5

The filter did not remove any snps, therefore why has the GQ score been downgraded?

Parameters as interpreted:

    --vcf test.vcf
    --recode-INFO-all
    --max-alleles 2
    --out test.test.vcf
    --recode
    --remove-filtered-all
    --remove-indels

After filtering, kept 1 out of 1 Individuals
Outputting VCF file...
After filtering, kept 4717895 out of a possible 4717895 Sites
Run Time = 156.00 seconds

Thanks!

A

Answer 1

[Brice Sarver]

Genotype qualities are calculated by setting the phred-scaled likelihood of the best genotype to zero and then calculating the difference from the next most likely. My guess is that your variant caller has a higher max GQ than is allowed by VCFtools (which might be capped at 99 by default) and VCFtools is placing it at the maximum. Not really a ton of difference between 99 and 123 - still quite confident in your call relative to the others.