SHRiMP alignment results for colorspace miRNA

[dsidote]

I'm trying to use SHRiMP to analyze a miRNA dataset using AB colorspace reads. I noticed when looking at the alignments using the pretty print option (-P) that SHRiMP aligns bases to the end of a read rather than the next base in the sequence. Does anyone know how to stop this behavior? I've tried various options but haven't figured out the solution yet.

Here is an example. the read and the reference are both ACTGTGGGCCCTTTCCGCACCA, yet SHRiMP aligns an A to the end of the read and calls it an insertion of AATCACCG.

>3_306_617 MIR-675Y_HSA-MIR-675___ - 1 22 1 30 30 129 12T8(AATCACCG)1
ACTGTGGGCCCTCTCCGCACC--------A 1
ACTGTGGGCCCTTTCCGCACCAATCACCGA
212111003002002033110103211032 30

[brentp]

That may be because it does a global alignment by default, you could try setting either the --local or --bfast flags to gmapper-cs and see if that helps.

[dsidote]

Aside from the read counts, I would also like to know what bases this particular enzyme adds to the end of the miRNA during amplification. I tried using the local setting and it does change the way SHRiMP reports the aligned reads, that is, its no longer reporting gaps where there shouldn't be, but it no longer reports the end of the read that does not align. I'll give the bfast option a try. Thanks.