A bug in the core pindel code is currently incorrectly calling complex indel regions. These calls have the same reference allele value and alternative allele value. They should not pass when using flags FF019 and FF020, but are currently passing due to the bug in the core code.
These calls have no value and were not called previously. Masking these incorrect results will not degrade the pindel results.
Keiran has suggested adding the following code:
if($record->ref_seq eq $record->alt_seq) { warn "SKIP: Dirty record due to issue in pindel core"; next; }
A bug in the core pindel code is currently incorrectly calling complex indel regions. These calls have the same reference allele value and alternative allele value. They should not pass when using flags FF019 and FF020, but are currently passing due to the bug in the core code.
These calls have no value and were not called previously. Masking these incorrect results will not degrade the pindel results.
Keiran has suggested adding the following code:
if($record->ref_seq eq $record->alt_seq) { warn "SKIP: Dirty record due to issue in pindel core"; next; }
within the following while loop:
https://github.com/cancerit/cgpPindel/blob/e574359b02055a6dea3e3b443502d63e1b1e1784/perl/bin/pindel_2_combined_vcf.pl#L195-L198
This will log the call as invalid and will exclude the record from the result vcf.
When a fix is released, the calls should then appear in the result vcf as expected.
Thoughts? @AndyMenzies @davidrajones