Samtools mpileup not providing correct count info

Question

[jgbradley1]

I simulated reads for a list of variant sites using GATK SimulateReadsForVariants tool. From that, I get a bam file output. Next I create a pileup using samtools mpileup.

samtools mpileup -t DP,DPR,DV,INFO/DPR -vuf GRCh37.fa -l snp.file.vcf simreads.bam > simreads.raw.vcf

The problem is that the output is not giving me correct counts at the variant sites. Here's an example of the first two SNP sites from the pileup.

#CHROM    POS    ID    REF    ALT    QUAL    FILTER    INFO    FORMAT    NA12878
1    837214    .    G    <X>    0    .    DP=20;I16=0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0;QS=0,0;MQ0F=0;DPR=0,0    PL:DP:DV:DPR    0,0,0:0:0:0,0
1    851390    .    G    <X>    0    .    DP=20;I16=0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0;QS=0,0;MQ0F=0;DPR=0,0    PL:DP:DV:DPR    0,0,0:0:0:0,0

So why would all the counts (with the exception of the DP tag) be 0? I opened up the bam file in IGV to check the first couple of sites and the reads do pileup there as they should.

[Devon Ryan]

Can you post the BAM file (or a small subset of it) and the version of samtools that you're using?

Answer 1

[jgbradley1]

I figured out the solution myself. I had simulated more error in my reads than I initially thought. For anyone else experiencing a similar issue, check the -Q, --min-BQ mpileup flag. Samtools automatically does not count reads in the other tags unless they are "high quality" (i.e. have a minimum base quality of 13). Depending on your experiment, if you're only interest is getting all the reads counted then consider using --min-BQ 1.