Mapper: Initial mappings (FASTQ -> BAM)

[eweitz]

In the Mapper module, support taking FASTQ input and converting to BAM output.

@chris-owen, @dauss75, this is one of the issues we discussed.

[chris-owen]

Segun and I produced this pipe using BWA mem and Samtools to map a PE fastq to the reference and produce a bam file that is sorted with duplicates removed.

bwa mem <reference_genome_fasta> <left_reads.fastq> <right_reads.fastq> | samtools view -Shu - | samtools sort - - | samtools rmdup -S - <PE_Mapped_DupRm.bam>

[dauss75]

snakemake is installed and currently putting some stuff in a test script to work with bwa and freebayes

[CarlosBorroto]

@chris-owen this looks really good, the only thing is I realized samtools rmdup will remove already marked duplicate reads. At least as far as I can tell. This means you need to mark duplicates first. The most common tool for this is picard's MarkDuplicates. However more recently I've seen samblaster as a better option as it is piping friendlier which picard is not.

Samblaster recommends doing something like:
bwa mem <idxbase> samp.r1.fq samp.r2.fq | samblaster | samtools view -Sb - > samp.out.bam

You should be able to integrate this into the command above pretty easily.

[CarlosBorroto]

Oh, and depending on the caller you might not need to remove duplicates just mark them. Callers these days are smart enough to ignore duplicate reads that has been marked as such.

[chris-owen]

Segun and I just met with Carlos and he confirmed the pipe commands above are sufficient.