When calculating consensus reads it happens that reads get skipped and directly written into a bam file.
As skipped records are expected to be used in downstream analysis it may happen that these reads are marked as duplicates while the remaining duplicates got merged into a consensus read.
In that case the marked reads would be ignored by variant callers.
To prevent this from happening reads the duplicate-flag will be removed from skipped reads and also the PicardTools specific tags: DI, DS and PG.
When calculating consensus reads it happens that reads get skipped and directly written into a bam file. As skipped records are expected to be used in downstream analysis it may happen that these reads are marked as duplicates while the remaining duplicates got merged into a consensus read. In that case the marked reads would be ignored by variant callers. To prevent this from happening reads the duplicate-flag will be removed from skipped reads and also the PicardTools specific tags:
DI,DSandPG.