chollenbeck / rad_haplotyper

MIT License
7 stars 5 forks source link

Added memory friendly haplotype builder #13

Closed chollenbeck closed 8 years ago

chollenbeck commented 8 years ago

Switched the haplotype builder based on Bio::DB::Sam to a new method based on Bio::Cigar. The new method uses samtools to output alignments in SAM format, and Bio::Cigar to identify the correct bases and build haplotypes.

The new haplotype builder also does not enumerate all possible haplotypes given the individuals genotype, which is computationally and memory intensive when there are many sites on a contig. Instead, all unique observed haplotypes are tested for whether they are possible or not, given the genotype.