why the vcf has so little high_quality bases mapping on the reference?

[wswill]

hi,all
Now,i have met a problem when using samtools to get SNP call. In the vcf file generated from samtools, i find high_quality bases so little in col(DP4=0,0,21,15), although coverage is very high(DP=5773). Then i check the sam file, and find that reads' BQ are almost high.
So,i want to know why .
The command:
mpileup -DugS -C 50 -f scaffold.fa all.bam |bcftools view -bcvg -> all.var.bcf
bcftools view all.var.bcf | vcfutils.pl varFilter > all.var.flt.vcf
The vcf file(partical):
gi|158420570|ref|NC_009938.1| 115 . C T,A 168 . DP=5773;VDB=0.0472;AF1=1;AC1=2;DP4=0,0,21,15;MQ=36;FQ=-132 GT:PLP:SP:GQ 1/1:201,105,0,183,71,180:36:0:99
gi|158420570|ref|NC_009938.1| 784 . G A 177 . DP=6382;VDB=0.0444;AF1=1;AC1=2;DP4=2,1,145,11;MQ=36;FQ=-282;PV4=0.21,9.3e-10,0.29,1 GT:PLP:SP:GQ 1/1:210,255,0:159:7:99

[tedtoal]

DP4 too small

I have the same problem that you described. DP4 values are small even though DP is high and quality is high. Did you ever find an answer to this?

[swbarnes2]

Quick thing to try, rerun mpileup with -B or -E. Sometimes, the BAQ calcautions will wrongly downgrade the heck out of quality scores at variations, make it hard for the caller to call them.

The other thing to do is to actually look at the .sam file of the reads that cross that locus. What is the mapping quality of those reads? What is the quality of the letter at that locus?

[Heisman]

Also include -A. That has solved this problem for me in the past.