using bwa to map illumina paired end reads

[mikeworth]

hello,

ugh... this is such a stupid question, i'm embarrassed to ask it, but i must. i can't find a definition of the workflow to use bwa to map paired end reads anywhere (am i simply missing it?). i'm starting with fastq seq files and i know how to index the genome. compiling info from the snooping i've done - this is what i believe the workflow to be:

$ bwa aln genome.index.fa s_4_1_sequence.fastq > s_4_1.sai
$ bwa aln genome.index.fa s_4_2_sequence.fastq > s_4_2.sai

$ bwa sampe genome.index.fa s_4_1.sai s_4_2.sai s_4_1_sequence.fastq s_4_2_sequence.fastq > s_4.sam

then, using samtools and s_4.sam i can create .bam and .bam.bai files.

is this correct? i think i understand the -b1 and -b2 options, but those are only used when starting w/ .bam files, which i'm not.

what is a little confusing to me is the two separate calls to 'bwa aln'. i suppose it's wildly wrong to (unix) cat the two s_4_[12]_sequence.fastq files together and make one call to 'bwa aln'?

thanks for your help,
mike

[swbarnes2]

I don't know that the sam file will reflect that you have paired reads if you cat the two fastqs together. It probably won't. If you carry out the default commands as you have written them, the software will figure out that the read are paired, and will set the flags, and ISIZE, etc correctly.

[mikeworth]

thanks for the help.

[mikeworth]

thanks. i've run bwa, as described above, w/ the paired end data and i believe everything went properly.

thanks again
mike

[cdry7ue]

Mikeworth,
Did you remove the sequencing adaptors or are you relying on BWA to do the soft clipping at the ends of the reads?

[mikeworth]

i removed the adapters (and more) from the reads. our data was barcoded and i separated the reads via a perl script based on the barcodes and removed everything up to and including the barcodes.

i'm curious why you're asking?

thanks,
mike

[jp.]

May please I get the commands to remove the seq-adaptor prior to bwa for PE Illumina ?