Vcf files fields: AD vs DP4

Question

[Rob]

In Vcf file, generate by samtools mpileup, there is an AD field (Allelic depth). This field is suppose to replace the previous DP4 field (Number of high-quality ref-forward, ref-reverse, alt-forward and alt-reverse bases).

But when I use -t DP4, AD in mpileup, I have a different number of values in the 2 fields:

GT:PL:DP:DP4:AD 0/1:27,0,107:5:4,0,1,0:4,1,0

And I don't understand why one data is missing in the AD field... Is it not supposed to be like the DP4, so:

high-quality ref-forward, ref-reverse, alt-forward and alt-reverse bases

?

After that I am not sure: what there are 3 values in the AD field ? And what are the differences between AD and DP4?

Thanks for your help.

Rob

Answer 1

[igor]

I think DP4 shows ref and all alt reads, but AD further separates the different alt alleles (a single record can have multiple alt alleles).

If you check the AD and DP4 descriptions in the VCF header, it should explain both fields.

I actually don't have AD in my samtools 1.3 VCF. They may have taken it out or it might depend on some other setting. Thus, I can't really confirm.

[Rob]

Hi, You are right, I thought there was one values missing in the AD fields, but it's not.

Here the description of the AD and DP4 field:

##FORMAT=<ID=DP4,Number=4,Type=Integer,Description="Number of high-quality ref-fwd, ref-reverse, alt-fwd and alt-reverse bases">
##FORMAT=<ID=AD,Number=R,Type=Integer,Description="Allelic depths">

The DP4 appears good, 4 values as expected. But for the AD field, why there is 3 values? The first appears to be for reference, and the second for allele, so, what is the third values for?

My complete vcf line is:

scaffold1124     33      .       A       G       63      .       DP=1018;AD=199,6,0;VDB=0.000411984;SGB=-8.52938;RPB=0.00019938;MQB=1;BQB=0.490025;MQ0F=0;ICB=0.00595507;HOB=0.00270329;AC=5;AN=136;DP4=199,0,6,0;MQ=60       GT:PL:DP:DP4:AD 0/1:27,0,107:5:4,0,1,0:4,1,0

Thanks again.

[igor]

The third one is 0, which makes sense as there is only one alt allele. Not sure why there is a value there at all. This is from a multi-sample VCF, so maybe one of the other samples is causing that.

[Rob]

Yes, it's probably that.. I think I got the things, thanks for your help!