BWA & Reverse Sequences

[dp05yk]

Hi all,

Just wanted to clear something up. When BWA reads the input sequences, it generates the reverse sequence for each read. Then, when searching for matches, it will check both the forward and reverse sequences. However, it will check them on the forward and reverse-complemented indexes, respectively.

I'm sure I'm incorrect here because this seems totally redundant. If we really wanted to check both strands of the reference sequence, why do we switch to the reverse sequence when switching reference strands? Would this give us the same result?

[dp05yk]

Oh, wait, now it looks like the sequence and reverse sequence (p->seq & p->rseq) in the code are actually identical unless they are SOLiD reads, in which case they are simply complements (but NOT reverse-complements).

It may have just been a case of weird variable names in the code, but some confirmation would still be nice!