Score Filtering

Score filtering

JBR supports BED-like and BigBed files filtering using an approach similar to Span Tuning section.

After you've annotated confident peaks and gaps you could filter BED-like track regions by score threshold which better matches annotations. By default fifth column is used for filtering but it is possible to change score column (e.g. containing p-value), see Region Score Column section.

To filter BED-like tracks:

• Enable Peaks Annotation Mode and Annotate confident peaks (see see Region Score Column section)

• Load BED-like / BigBed track (see Load Tracks section)

  

• Select tracks and choose Tune BED Tracks context menu action.

  

• After progress has finished you will get the result:

  

  You could change filtering threshold manually using Filter by Score.. menu action after filtering has been completed.

  

• Export filtered tracks to BED files using Export Tuned BED Tracks Peaks... context menu action.

  

About Track

Filtered tracks About Tracks results include additional information about mismatched annotations number for filtered track:

============ Track (1 / 1): 'SICER H3K36me3 BB' ============
Bed Format: (bed4, '\t')
Track source: https://artyomovlab.wustl.edu/publications/supp_materials/SBW-epigenetics-StL/data/GSE112622/chipseq/peaks/sicer/GSM3074494_BB_GVO_H3K36me3_mm9.scoreisland
Source size: 1,5 mb
Peaks Statistics:
  Peaks number: 40 604
  Peaks summary length : 410 987 146 bp
  Genome (mm9 [all chromosomes]) coverage: 0.15%
  Lengths:
    Min: 599 bp
    Mean: 10 121 bp
    Max: 216 449 bp
    5%: 3 299 bp
    50%: 7 349 bp
    95%: 26 249 bp

  Track error score:
    Total error: 31% (13/42)
    No peaks error: 0% (0/12)
    Peaks error: 20% (2/10)
    Peak start error: 50% (5/10)
    Peak end error: 60% (6/10)