I'm trying to automatically compare BAM files being output by bowtie2 (for a continuous integration system). Currently using bamUtil diff
which works well, except I'm noticing sometimes 2 lines in the BAM file will have the same name/start position, but represent different reads. For whatever reason, bowtie2 sometimes reverses the order of the lines, so bamUtil's algorithm calls a mismatch. I see several potential solutions to this problem:
Get samtools to sort by another dimension in addition to position/name. Pretty much anything should do. Can this be done? I don't see any options in the man page.
Use another comparision tool than bamUtil which doesn't have this problem. Honestly a stand-alone program for diffing would be best for me anyway and I'm considering writing one.
Find some way to get bowtie2 to be more deterministic in its ordering. Any ideas here?
show us an example please.
I confirmed that bamHash shows the files as matching. I think my solution for now is to just use that. I'm still a little concerned that I don't know which stage of my pipeline is interleaving the reads. From what I've read I should be accounting for all sources of non-determinism (using single core, etc). I might dig deeper into this at some point but after reading up on bamHash and testing it out I think that's what I need. Thanks for the help. If you move your comment suggesting bamHash to an answer I'll accept it.
I've moved my comment to an answer. Please double check, though, that BamHash is really doing the comparison you want and that you don't care about mapping positions being the same (if you just want to ensure the same entries are contained in both files, regardless of whether/how they're mapped then you're good to go with BamHash).
Now that I understand the problem better, you may actually be right about mappings being important for my application. Read names/sequences might be enough of a guarantee for us, but I'm waiting to hear back on that.