Coverage visualization; a companion viewer for indexcov results.

Here we use indexcov to quickly estimate the coverage across samples then find regions of large, coverage-based anomalies. The aim is to highlight regions of significant (passing the user's z-score threshold) and sustained (beyond user specified distance) deviation from the majority of samples. Significance is determined using z-scores (--zthreshold) for all samples at all points using median absolute deviation, but in order to be highlighted, points must be significant consecutively throughout a user specified distance (--distancethreshold).

Install nextflow:

curl -s https://get.nextflow.io | bash

With Docker or Singularity:

nextflow run brwnj/covviz -latest -profile docker \
    --indexes 'data/indexes/*.crai' \
    --fai data/g1k_v37_decoy.fa.fai \
    --gff data/Homo_sapiens.GRCh37.82.gff3.gz

Which gives us ./results/covviz_report.html.

• --indexes
  • quoted file path with wildcard ('*.crai') to cram or bam indexes
• --fai
  • file path to .fai reference index
• --gff
  • file path to gff matching genome build of --indexes

• --outdir
  • output directory for results
  • default: "./results"
• --sexchroms
  • sex chromosomes as they are in --indexes
  • default: "X,Y"
• --exclude
  • regular expression of chromosomes to skip
  • default: "^GL|^hs|^chrEBV$|M$|MT$|^NC|random$|Un|^HLA\-|_alt$|hap\d+$"
• --zthreshold
  • a sample must greater than this many standard deviations in order to be found significant
  • default: 3.5
• --distancethreshold
  • consecutive significant points must span this distance in order to pass this filter
  • default: 150000
• --slop
  • leading and trailing segments added to significant regions to make them more visible
  • default: 500000
• --project
  • can be used to name your indexcov to something more meaningful
  • default: "NF"

See: https://brwnj.github.io/covviz/

Significant regions will be displayed in color atop a gray region which represents the upper and lower bounds of a given point minus any values deemed significant.

The metadata table will be displayed below the plots.

Clicking on plot traces highlights the line and searches the metadata. Double-clicking de-selects lines, resets the plot, and de-selects samples from the table. Clicking on the gene track launches a search for the gene's respective Gene Card. In cases where genes overlap, multiple windows/tabs will be opened.

covviz is free and unrestricted for non-commercial use. For commercial use, please contact [[email protected]].