Bowtie alignment of piRNAs to find potential targets

Question

[mujupas]

Hi, I have a set of piRNAs in fasta format and I want to find their potential targets. According to literature nucleotides in position 1-10 in the piRNAs sequences should constitute the seed and do not tolerate any mismatch while nt 11-20 tolerate up to 3 mismatches. Bowtie can specify the length of the seed, but I'm having troubles finding the right combination of parameters to translate the piRNA target pairing logic into a single bowtie command line.

Is there any solution to get this done with a single alignment run or do I have to run a first alignment guided by the seed 1-10nt, and successively another alignment for the 11-20 region allowing for up to 3 mismatches?

Thank you in advance for your help!

Answer 1

[Charles Plessy]

How about running the alignment tolerating 3 mismatches anywhere, and then filtering the results using the CIGAR string ? (You would need an aligner that produces CIGAR strings distinguishing matches (=) and mismatches (X); I do not remember if bowtie does).

Answer 2

[Charles Plessy]

Or maybe you can use LAST with a custom seeding scheme ?

Answer 3

[mujupas]

Thank you, I'll try both options!