Output Section

Output section #

[output]
    directory = "output" # Where do we place the output files?
    write_annotated_bam = false # if set to true, write <directory>/annotated.bam
    mode = Region|SingleCell|StartPositions|Coverage|None # optional, see below.

Controls where we place the output.

Output modes #

If you leave mode off, it will default to either single cell quantification (if cell_barcodes are present) or region quantification (if no cell_barcodes are present). You can overwrite this if you want to, for example, deduplicate per cell, but count per region.

Region #

  • a counts.tsv file with the counts per region, (with columns being <region_id> count_correct count_reverse
  • a `counts.tsv.stats.tsv’ with some summary statistics

The region id comes from your input definition. For GTF, either the aggr_id_attribute or the id_attribute is used, for references it’s ‘reference’, and for BAM_tags it’s the two letter tag.

SingleCell #

(if a [cell_barcodes] section is present)

StartPositions #

  • a start_positions.tsv file with the count of each (corrected) read start postion for both strands.
  • a start_positions.tsv.stats.tsv with some summary statistics

Positions are in genomic coordinates, and 0 based.

Only reads that ‘hit’ regions are counted. Perhaps use bam_references as your region source if you want to count them all.

Coverage #

  • a coverage.tsv file with the coverage of each detected position. Two count columns, so accounting for both strands.

Positions are in genomic coordinates, and 0 based.

Only reads that ‘hit’ regions are counted. Perhaps use bam_references as your region source if you want to count them all.

None #

Don't create count outputs.

annotated.bam #

If requested, we output the decisions on each read as <directory>/annotated.bam.

We add the following tags (and remove their old values if they were set in the BAM file):

Note that depending on where exactly reads are filtered or detected as duplicates, some of the tags may not be set.

XF:i - filter decision #

  • 1 - the read was removed by a filter
  • 3 - the read was detected as an UMI duplicate
  • 4 - the read’s cell barcode was not in the whitelist
  • 5 - the read had no barcode
  • 6 - the read had no UMI
  • 7 - the read was an approximate UMI duplicate
  • counted reads do not get an XF tag.

XQ - correct hits #

Genes (regions) hit in the correct orientation, comma separated

XR - reverse hits #

Genes (regions) hit in the incorrect orientation, comma separated

XP - corrected position #

The corrected position of the read. See TODO. Only present if input.correct_reads_for_clipping is set to true.

CB corrected cell barcode #

The (corrected) cell barcode.

CR uncorrected cell barcode #

If a read’s barcode was not in the white list, the uncorrected barcode is stored here. (otherwise this tag is not set)

UMI CR / UR #

TODO (not yet implemented)