Mutect1.1.7 clustered position filter for soft-clipped reads
Please can you consult me.
I have an example of a candidate somatic variant where all variant reads are soft-clipped (3 bases from the mutation).
The variant is naturally rejected by Mutect when running with default parameters.
However, disabling clipread filter doesn't help and the failure reason is "clustered position".
So the question is: if the caller does not distinguish between the start/end of the read and the start/end of the alignment?
I got curious and will appreciate any input if some of you came across similar issues.
(PS. I do realize that aforementioned filters are generally reasonable and the reason I am experimenting is that the mutation in question is clinically relevant )