Hybrid reads

[veena]

Hi,

I want to align nextgen seq reads that should only partially align to one reference seq. I'm most interested in the part of the read that doesnt align to the reference, but need the aligned portion to filter for it. What aligner would work best? BFAST? I'm working on a Mac OSX to start.

Thanks much!
-V

[swbarnes2]

It's very old fashioned, and probably too slow for mammalian genomes, but SOAP has an option to iteratively trim from the 3' end until the remnent aligns well enough.

From the SOAP output, you could get the names of reads which had to be severely trimmed to ailgn, and how far they had to be trimmed. Then pull out those reads, and look at the sequence after the aligned portion.

[frozenlyse]

Another option is to use Bowtie, and iteratively map data that doesnt align clipping the 3' or 5' end of the read (whichever end you want to remove) a few times... hopefully bowtie is fast enough that mapping over and over isnt too much of a problem