Why different read depth on IGV and in VCF for a variant

Question

[pragnapcu]

For a variant, the read depth in the VCF file (hardfiltered) is 10 whereas, on IGV, the total count is 79. What might be the reason?

Answer 1

[LChart]

Both the variant caller and IGV have default thresholds for mapping quality; could it be that the variant caller was run with a higher mapping quality threshold, and that many of the reads viewed in IGV fall below that threshold? (Also, depending on the caller, the DP field may reflect the filtered or raw depth).

[pragnapcu]

Yes. It makes sense but, how can I validate the variant in this case (as the depth is low)?

[Friederike]

You'll have to re-sequence that locus. Or look at replicate samples that are expected to have the same variants.