yet another question regarding vcf format

[mstagliamonte]

OK, I am sorry to ask yet another question regarding vcf format, I have noticed there are similar questions asked everywhere, on any forum. Still, I have a doubt, and wonder if I am misunderstanding some of the tags.

I have done some variant calling using samtools 1.2;
The format field is structured as follows:

##FORMAT=<ID=PL,Number=G,Type=Integer,Description="List of Phred-scaled genotype likelihoods">
##FORMAT=<ID=DP,Number=1,Type=Integer,Description="Number of high-quality bases">
##FORMAT=<ID=DV,Number=1,Type=Integer,Description="Number of high-quality non-reference bases">
##FORMAT=<ID=DPR,Number=R,Type=Integer,Description="Number of high-quality bases observed for each allele">
##FORMAT=<ID=SP,Number=1,Type=Integer,Description="Phred-scaled strand bias P-value">
##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">

I understand the difference between FORMAT/DP and INFO?DP (the latter considers all the bases), but I recently realized that I may be misunderstanding the difference between FORMAT/DP and FORMAT/DPR tags.
Most of the times, the sum of the values in DPR is equal to the DP; in some cases though I have noticed that this is not true, and the DP value is slightly larger. In these cases, I have noticed that the alternative allele coverage in DPR also differs from the one reported in DV, by the same amount.

Can you please help me understanding what I'm missing?

Thanks,
Max

EDIT:
Here is an example:
Format= GT : PL : DP : DV : SP : DPR
Sample= 1/1 : 255,255,0 : 251 : 248 : 5 : 3,244

[westerman]

I am not sure but I will make the, possibly unhelpful, suggestion to use Samtools 1.3 instead of 1.2. It may be more accurate and useful. From the change log:

The mpileup command now outputs the unseen allele in VCF/BCF as <*> rather than X or <X> as previously, and now has AD, ADF, ADR, INFO/AD, INFO/ADF, INFO/ADR --output-tags annotations that largely supersede the existing DV, DP4, DPR annotations.

[mstagliamonte]

Thank you, Westerman,

That's something I didn't think of. So basically, the missing bases support an allele that was not called, am I correct?

[westerman]

That is my guess. Not 100% sure though.

[mstagliamonte]

OK,

Thank you for your help