Bioinformatics Stack Exchange is a question and answer site for researchers, developers, students, teachers, and end users interested in bioinformatics. It only takes a minute to sign up.
Sign up to join this community
Is it worth mapping bisulfite reads (WGBS) to a reduced size genome?
I'm interested only in modifications in CpG context, thus instead of mapping to a whole genome (human genome) I would:
My idea is to reduce mapping time, however afraid that there might be some mapping bias or other issues.
CG sequences are pretty common; I expect you'll hit almost all of the genome when looking at a 2kb window centred on every CG, so there wouldn't be much point in doing that for efficiency purposes.
dariober has pointed out that reads coming from elsewhere in the genome could be incorrectly mapped to the reduced genome with high mapq score since the best, genuine genomic region is not present. So even if you are interested in few small regions, you should still map to the entire genome.
Don't restrict the portion of the genome you're mapping to, you'll just increase false-positive alignments by doing so. If you're worried about mapping time, then (A) use more cores or (B) use a different mapper (BSMAP was the quickest out there last I looked, though in my opinion it's fine to wait a bit longer for better quality results (I normally suggest bwameth to people)).
As pointed out by user 719016 in the comments, the above does not apply to amplicon sequencing. With that, you'll get VASTLY better performance if align to a reduced-size genome...and without an appreciably increased false-positive rate (unless your target-enrichment for sequencing is poor).
