Questions about mapping paired-end data by BWA

Question

[Proxy Agent]

Hello everyone, I am confused about a strategy hired to mapping paired-end sequencing data which was employed in the paper Accurate identification of human Alu and non-Alu RNA editing sites. In brief, they mapped each of the paired-end reads separately using the commands “bwa aln fastqfile” and “bwa samse -n4”, and then merged these reads into a single file. But bwa has the command bwa sampe, why they just not use that command?

[Matteo Schiavinato]

It looks weird, not to make use of the paired information. Probably it has to do something with the fact that they are studying editing, and therefore there can be a difference between reads and template.

Answer 1

[Arup Ghosh]

Previously BWA didn't have support for direct alignment of paired-end sequence data. So, first, you had to align the files in .sai(suffix array index) format separately then merge the matching pairs.

Now, with bwa aln -1 <read1.fq> -2 <read2.fq> you can directly align paired end reads.