Can a SNP with 35 vs 54 reads really be homozygous?
We have sequencing information for a whole family (father, mother, 3 kids) and analyzed them with the gatk (more info below). And while looking through the results I found a SNP with these predicted genotypes and read frequencies:
kid 1: 0/0, 168 vs. 26 reads
mother: 0/0, 211 vs. 34 reads
father: 0/0, 220 vs 18 reads
kid 2: 1/1, 35 vs. 54 reads
kid 3: 1/1, 20 vs. 42 reads
First it caught my eye because of the impossible heredity but than I also noticed the, at least to me, strange genotype calls for kid 2 and 3. Can you explain the effects that lead to such results?
exact vcf entry:
4 71347185 . T C 332.95 . AC=4;AF=0.400;AN=10;BaseQRankSum=3.959;DP=828;Dels=0.00;FS=0.000;HaplotypeScore=11.5830;MLEAC=4;MLEAF=0.400;MQ=29.13;MQ0=153;MQRankSum=-6.418;QD=2.20;ReadPosRankSum=-10.775;SOR=0.708 GT:AD:DP:GQ:PL 0/0:168,26:194:99:0,157,2194 0/0:211,34:245:93:0,93,2661 0/0:220,18:238:99:0,101,2422 1/1:35,54:89:15:131,15,0 1/1:20,42:62:27:250,27,0
pipeline (not all under my control):
- mapper is unknown, probably bwa
- Picard MarkDuplicates (Version 1.128)
- GATK IndelRealigner (Version unknown)
- GATK UnifiedGenotyper (Version 3.4-46), I know it is deprecated but we had to use it for a reason