Samtools Vcf Format Depth Discrepancy?

Question

[Zev.Kronenberg]

Here is a VCF line directly from the Samtools variant calling pipeline (no varfilter.pl) for one exome:

chr1    156713512       .       C       T       60.5    .       **DP=6**;AF1=1;AC1=2;DP4=0,0,4,0;MQ=34;FQ=-39       GT:PL:**DP**:SP:GQ  1/1:93,12,0:**4**:0:21

I noticed that the genotype column has a depth of 12 listed (highlighted: **). How can this be if DP=6? What am I missing?

Answer 1

[Jeremy Leipzig]

That 12 is not DP, DP is 4 (high quality bases)

GT:1/1
PL:93,12,0
DP:4
SP:0
GQ:21

[Zev.Kronenberg]

Sorry, I accidentally highlighted the wrong field. How are high quality reads defined? Seems odd not to count the total reads per individual...

Thanks

[Jeremy Leipzig]

I think phred >=20

Answer 2

[Damian Kao]

In the [?]VCF specs[?]:

DP in info column (8th):

DP - combined depth across samples, e.g. DP=154

DP in genotype column (9th):

DP - read depth at this position for this sample (Integer)

One is for across samples and other is for this sample only.

Answer 3

[Deniz]

How did you get this vcf output? dafault setting gives GT:PL:GQ format. Could you write your command?

Thanks

[Zev.Kronenberg]

Pretty much used the defaults from SAMtools why?

[Deniz]

Your format is different. That's why I wondered. Are you happy with results and format?