Variants with multiple functional annotations -- counting occurrences

Question

[olavur]

I annotated my VCF with SnpEff, whether they are missense variants, intron variants, and so on. Now I want to know how many variants of each different type I have. I don't know exactly how to do this because each variant may have (and often has) several functions (usually corresponding to several overlapping genes, as far as I understand).

Considering this, how would one usually go about counting the number of variants of each functional type?

Answer 1

[Kevin Blighe]

I don't know exactly how to do this because each variant may have (and often has) several functions (usually corresponding to several overlapping genes, as far as I understand).

Any given variant can indeed overlap multiple genes that may, for example, share the same promoter; however, each variant can also have different functionality within the same gene's transcript isoforms. A variant can be exonic in one isoform, for example, but intronic or splicing in another.

SnpEff does output a summary file: http://snpeff.sourceforge.net/SnpEff_manual.html#outputSummary

Variant Effect Predictor also outputs a HTML stats file, which may be of use to you:

[olavur]

I see, thank you. I wonder, though, if SnpEff counts all annotations of all variants or just the first one in the list (which is sorted by impact in decreasing order).

[Kevin Blighe]

It appears to be an accumulation of all noted annotations. The author attempts to explain it on the page to which I linked you (above), but it's not something that is easy to explain!

The easiest way to test this is to just annotate a handful of variants in your VCF, and then manually check the annotated stats / results.