shabam

A python tool to plot BAM or CRAM sequence reads.

Installation

install cairo if not already installed:

# macOS via conda, or via homebrew (choose one, then set the library path)
conda install cairo; CAIRO=`conda info --root`/pkgs/cairo-1.14.8-0/lib
brew install cairo; CAIRO=`brew --prefix cairo`
export DYLD_LIBRARY_PATH=${CAIRO}

Install shabam:

pip install git+git://github.com/dlrice/shabam.git --user

The dream

from shabam import seqplot
seqplot('example.bam', chrom='1', start=30243, end=30321,
    fastafile='reference.fasta', out='plot.svg')

Plotting options

• shade reads by strand with by_strand=True
• plot multiple sequence files together with a list of paths e.g. ['child.bam', 'mom.bam', 'dad.bam']
• export PDF, PNG, SVG or PS formatted plots with matching filename extensions
• seqplot returns PNG data if you don't include an output filename

Command line version

./bin/shabam.py \
  --seqfiles example.bam \
  --chrom 2 \
  --start 30243 \
  --end 30321 \
  --fastafile reference.fasta \
  --out plot.svg

Further improvements to the dream

• [ ] Use consensus sequence when we don't provide a reference sequence
• [ ] Allow custom colors
• [ ] Compute proportion of variants at any site
• [ ] Plot read depth
• [ ] At positions with proportion of variants > threshold, reflect proportion with base colors
• [ ] Option to scale plotted base size, currently at 10 pixels per base
• [ ] Optionally shade plots deepVariant style:
  • red channel: nucleotide
  • blue channel: read strand
  • green channel: base quality
  • alpha: base supports ref or alt
• [ ] Flatten vertical plotting in high depth sequence data
• [ ] Down-sample reads for extremely high depth sequence data (>1000X)

Credit

Initial cigar parsing code lifted with permission from pybamview.