Can anyone tell me why BCFTools Call would call a homozygous variant at this position with 1/2 alt variant reads?
0
0
Entering edit mode
5.0 years ago
fwuffy ▴ 100

Here's a variant SNP with raw depth of 2, and DP4 field shows 1 alt-forward read. Can someone help me understand why this is called homozygous 1/1 at this position intead of hetero 0/1? (I am filtering out calls with DP < 5 after this, but it still raises questions)

Thanks!

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


chr19 3136246 . A G 14.1809 . DP=2;SGB=-0.379885;MQ0F=0;AF1=1;AC1=2;DP4=0,0,1,0;MQ=60;FQ=-29.9903;ANN=G|5_prime_UTR_variant|MODIFIER|GNA15|GNA15|transcript|NM_002068.3|protein_coding|1/7|c.-205A>G|||||205|,G|downstream_gene_variant|MODIFIER|LOC390876|LOC390876|transcript|TRANSCRIPT_LOC390876|pseudogene||n.*4809A>G|||||4809| GT:PL:DP 1/1:44,3,0:1

Edit: I may have answered my own question, in that there should be a value in the ref fwd/ref reverse DP4 positions to make a heterozygous call, which are probably missing due to low quality, but I'll leave this up in case people have something to add...

like "Sum can be smaller than DP because low-quality bases are not counted."

bcftools vcf variants • 2.1k views
0
Entering edit mode

Yes, you have answered your own question. DP is the total reads at the position, whilst DP4 relates to 'high quality' reads.

Still, it is suspect to call homozygous from just a single read and, just to be sure, do not in any way base any clinical decision on this variant ;)