dodi chooses an optimal set of alignments from a candidate list. This can be used for improving split-read mapping, or for selectively mapping to target regions of the genome
Install using:
$ git clone https://github.com/kcleal/dodi.git; cd dodi $ pip install .
Requires Python>=3.6, cython and >=c++11 compiler.
Given a list of candidate alignments in .sam format, dodi will select an optimal spanning set of alignments. Common usage is to run dodi in a pipe, downstream of a mapper such as bwa mem or Last.
To use with bwa mem, make sure the -a flag is used, to report all alignments:
$ bwa mem -a -t8 ref.fa read1.fq read2.fq | dodi - > output_alignments.sam
To selectively map to target regions of the genome, supply dodi with a .bed file containing the target regions, and a --bias value (default is 1.15). Alignments falling within target regions receive a temporary bias to their alignment scores, meaning those alignments are more likely to be chosen by dodi:
$ bwa mem -a -t8 ref.fa read1.fq read2.fq | \
dodi --include target_regions.bed - > output_alignments.sam
For long-read mapping, a lower --bias value is recommended, e.g. 1.01