BurrowsWheelerAligner

Minimal WIP bindings for bwa (https://github.com/lh3/bwa)

Warning : make sure you use the same version of BWA as https://github.com/JuliaBinaryWrappers/BWA_jll.jl (0.7.17) to generate the reference index.

Usage

Single read alignment

using BurrowsWheelerAligner
const BWA = BurrowsWheelerAligner
import BWA.FASTA

# this folder include all the files generate by `bwa index`
index_file = joinpath(@__DIR__, "data", "genome.fa")

aligner = BWA.Aligner(index_file)

record = FASTA.Record(""">test\nGAGTTTTATCGCTTCCATGACGCAGAAGTTAACACTTTCGGATATTTCTGATGAGTCGAAAAATTATCTT""")
aln = BWA.align(aligner, record)[1]

@assert BWA.position(aln) == 1
@assert BWA.cigar(aln) == "70M"
@assert BWA.mappingquality(aln) == 60
@assert BWA.is_rev(aln) == false
@assert BWA.refname(aln, aligner) == "PhiX"

Paired-end alignment

In paired-end mode align will return SAM records (and return only the primary alignement):

    aligner = BWA.Aligner(index_file; paired = true, nthreads = 8)

    r1 = FASTA.Record("test1", "TGCGTTTATGGTACGCTGGACTTTGTGGGATACCCTCGCTTTCCTGCTCCTGTTGAGTTTATTGCTGCCG")
    r2 = FASTA.Record("test1", "AAAGGCAAGCGTAAAGGCGCTCGTCTTTGGTATGTAGGTGGTCAACAATTTTAATTGCAGGGGCTTCGGC")
    
    sam1, sam2 = BWA.align(aligner, (r1,r2))

Multithreading can be used by providing an array of paired records :

    using Random
    records = Tuple{FASTA.Record,FASTA.Record}[]
    for i in 1:1000
        n = randstring(rand(10:20))
        r1 = FASTA.Record(n, "TGCGTTTATGGTACGCTGGACTTTGTGGGATACCCTCGCTTTCCTGCTCCTGTTGAGTTTATTGCTGCCG")
        r2 = FASTA.Record(n, "AAAGGCAAGCGTAAAGGCGCTCGTCTTTGGTATGTAGGTGGTCAACAATTTTAATTGCAGGGGCTTCGGC")
        push!(records, (r1,r2))
    end

    alns = BWA.align(aligner, records)

Additional methods

The aligner mode can be switched with :

BWA.flag!(aligner, BWA.LibBWA.MEM_F_PE) BWA.flag!(aligner, BWA.LibBWA.MEM_F_NOPAIRING)