unmapped reads

[rjovelin]

debarcer is not currently filtering out unmapped reads. This introduces a bias in QC plots and artificially inflate read depth per umi group.
It seems logical to filter out unmapped reads, and keep the aligned mate (when appropriate) since debarcer is not using information about paired reads during grouping or collapsing.

[rjovelin]

Unmapped reads are skipped during Grouping. Collapsing already loops over pileup objects which only include mapped reads. The number of mapped and unmapped reads per genomic interval is recorded in the Stats directory, which are then used for QC plots.