Indel calling #91
Description
Is your feature request related to a problem? Please describe.
Not a clearly defined problem, but we want to gain confidence in INDEL calls.
Describe the solution you'd like
The solution is still unclear, but we would like to make a choice of a reliable indel calling strategy.
This would imply testing VarDict pileup mode, vs VarDict calling mode vs PINDEL calling vs any others that we might want to consider.
Describe alternatives you've considered
For now we are calling them in the same way as the SNVs with VarDict pileup but we are not so sure how much we can trust it.
Additional context
Some testing already ongoing with VarDict calling, but we should just switch the focus to the indels, and regarding pindel @FerriolCalvet tried it with some normal bladder samples, but there was no clear outcome of the benchmark since it was done in a single sample.
See here for more details on the testing:
/data/bbg/nobackup2/prominent/bladder/2024-05-17_deepUMIcaller_allsamples_newasminusxs/running_pindel/
see here for details on the comparison of indel calls:
/data/bbg/projects/bladder_ts/notebooks/complete_cohort/2025-01-02_pindel_output_processing.ipynb