Hello, I am wondering why, if I have used the -v option in bcftools (output only variant sites), I have in the output some reference homozygotes. Could it be because the genotype likelihood is too low that could be also a het? Does the -v option look at the most probable genotype (the one written down) or at the genotype likelihoods?

bcftools view -Svc myfile.vcf.gz

#CHROM  POS     ID      REF     ALT     QUAL    FILTER  INFO    FORMAT  MET_03912_4_1_TAGCTT
[...]
1       15906   .       A       .       3.55    .       DP=3;QS=0.000000,0.906977,0.093023,0.000000;VDB=4.034274e-02;AF1=1;AC1=2;DP4=0,0,0,3;MQ=16;FQ=-33;AF1=1;AC1=2;FQ=-30    PL:DP   31,3,0:3