VCF PL Values not making sense

Question

[fwuffy]

Hi- Can someone help me identify these PL (phred likelihood) genotype values for this multi-allelic indel row in a VCF?

The VCF row:

chr1    36418926    .   caagaaa cAAAAaagaaa,caa,cAAaagaaa   9.95049 .   INDEL;IDV=4;IMF=0.666667;DP=6;VDB=0.0340507;SGB=-0.556411;MQSB=1;MQ0F=0;AF1=0.502508;AC1=1;DP4=1,0,3,1;MQ=59;FQ=-14.6836;PV4=1,1,0.342519,1 GT:PL:DP    0/1:72,25,45,27,0,47,40,13,15,59:5

With 3 alt alleles, I expect 6 values, not 10

A/A:72, 
A/B:25, 
B/B:45, 
A/C: 27, 
B/C: 0, 
C/C: 47, 
??: 40, 
??: 13, 
??: 15, 
??: 59

These are unphased, one sample. called with bcftools 1.6+htslib-1.6 Using this to call variants:

samtools mpileup -v -u -B -t DP -f full.fa mapped-sorted-mkdup.bam | bcftools call -c -v - > germline-variants-all.vcf

Thanks

Answer 1

[swbarnes2]

You have the reference allele and 3 alt alleles, so there are 10 possibilities.

That locus is very poorly covered, and those alignments might not be right. I'd either eyeball it, to see if you can figure out what's going on there, or ignore it, on the grounds that it's too poorly covered to assess.

[fwuffy]

Oh right!!! Duh. It's been a while...

so I'd have

A/A:72, 
A/B:25, 
A/C:45, 
A/D: 27, 
B/B: 0, 
B/C: 47, 
B/D: 40, 
C/C: 13, 
C/D: 15, 
D/D: 59

The coverage is low cause it's a sub-sampled fastq. I also know there are issues around indels but just for the sake of knowing what i'm supposed to be looking at, setting aside the correctness of it for now.

*edit: This also confuses me cause the genotype call is 0/1.

Thanks.