#4 Validation step loses indels in short exons
Hi again,
Been testing Scalpel on some cell line data which have an EGFR deletion in exon 19 (AGGAATTAAGAGAAGC > A). The exon itself is fairly short, 55242361-55242527 in the capture file. The deletion is reported in single sample mode and in the main folder of somatic mode, but not in the validation folder. I tracked this down to two parameters (which both independently cause no indel being reported): increase of --radius from 100 to 150 in validation (probably going outside the exon boundaries) and dropping --covratio from the default 0.01 to 0.0 (is this intentional?)
I noticed that around --covratio 0.01 the runtime is very low but for changes from 0.01 (to, say 0.05 or 0.001) the runtime increases quite a lot and no indel gets reported for my test data.
Any pointers/comments/feedback welcome!
Can you please attach or email the portion of the BAM file for the exon containing the deletion? Thanks.