bwa map single and PE reads to ref?

[AdrianP]

Hello,

I want to store detect my variants in a VCF format. For that I need to map my PE reads (100x2) to reference. What I have discovered however is that for a lot of my PE reads, the fragments size is smaller than 200bp, which means some PE reads overlap, increasing the coverage of a region artificially.

I thought I would merge the PE reads with SeqPrep to place all PE reads that overlap in a merged.fastq file.

My question is, how do I then map both unmerged PE reads and merged SE reads to the genome with bwa? I only know how to do it if it's only SE or only PE.

Also, does this approach sound reasonable?

Adrian

Biostar: http://www.biostars.org/p/97423/#97430

[Brian Bushnell]

Map the single-ended reads, then map the paired reads, then merge the two sam files. We do that all the time.

[AdrianP]

Sorry I have never merged 2 sam files together, how would I do that?

cat *.sam > new.sam

?

Thanks for the advice!

[Brian Bushnell]

Try samtools merge.

[dpryan]

As I mentioned on Biostars, just use a variant caller that isn't confused by paired-end reads. GATK is one such example.