How GATK pipeline called a homozygeous alternate allele (1/1) although one copy is supported by 0 reads ?

Question

[mohsamir2016]

I have annotated SNPs using GATK, and have observed that in many 1/1 SNP sites (homozygeous alternate) the whole read depth (stated in DP) are supporting the the alternate allele (as seen in the AD field), while the reference allele is supported by 0 reads. So how GATK called this ref allele if there is no reads supporting it.

Example:

[Ram]

I've removed irrelevant tags from your post.

[Jeremy Leipzig]

So how GATK called this ref allele if there is no reads supporting it

it didn't call it reference. it called it hom_alt. where in your figure do you think it called it reference?

Answer 1

[Ram]

Ref alleles are from the reference genome. GATK doesn't need a read supporting the ref genome on a sample to know what the ref base is at a location.

[mohsamir2016]

Sure, but n my case 1/1 the allel is different from the reference allele and it called it as T, for instance. So it needs reads in the sample to support that. Also i have seen for many of 0/1 genotype a read for the reference allele ?

[Ram]

I don't understand what you're saying. Your allele is different from the ref allele so it needs the sample to support that? Why?

Give me the following details please:

1. Organism
2. Genome version used
3. Chr and Position
4. Ref allele
5. Alt allele

[mohsamir2016]

• Organism: gallus gallus
• Genome version used: GRCg6a
• Chr and Position: NC_006113.5, 24618
• Ref allele: A
• Alt allele: G
• Genotype: 1/1 (G/G)

So my understanding is that there should be sequenced reads (from ones sample) that support the two copy (allele) of the position. and the n of these reads is written in the field AD in VCF file. Also the field DP is the total read depth supporting that variant. Do you mean that because it is 1/1, so both copies is different from the reference and that is why the 0 in the AD is because no reads is saying that any of these copies at that position is similar to the reference?

[Ram]

Do you mean that because it is 1/1, so both copies is different from the reference and that is why the 0 in the AD is because no reads is saying that any of these copies at that position is similar to the reference?

Yes, you nailed it.