SomaticSniper VS Varscan2

Question

[user230613]

Hi guys,

As far as I've read, both SomaticSniper and Varscan variant calling algorithms have been developed by the same group. Do you have any experience of which one is more accurate? Or in which conditions should I use one or the other?

Thank you in advance,

Answer 1

[Chris Miller]

It just depends on what you're doing. Somatic Sniper has largely been superceded by a new generation of callers (Mutect, Strelka, etc) but is still useful in the niche case of having tumor contamination in your normal sample (very few other callers are designed to handle this case). In general, VarScan has kept up with the times pretty well.

That said, relying on a single caller is probably not your best bet. Each has strengths and weaknesses, and an ensemble approach followed by intelligent merging is probably your best bet for maximizing your performance.

[user230613]

Thank you very much! However I don't get you point about the advantages of using SomaticSniper when tumor contamination is expected in normal sample. As far as I can read in Varscan2 help, Varscan2 can handle this issue with the following two arguments:

--normal-purity - Estimated purity (non-tumor content) of normal sample [1.00]
--tumor-purity - Estimated purity (tumor content)

Am I missing something?

Answer 2

[TriS]

there are a number of papers out there comparing them (with other variant callers). examples are:

Evaluation of Nine Somatic Variant Callers for Detection of Somatic Mutations in Exome and Targeted Deep Sequencing Data Detecting somatic point mutations in cancer genome sequencing data: a comparison of mutation callers A comparative analysis of algorithms for somatic SNV detection in cancer