Looking for an aligner allowing intron gap

[superligang]

We are mapping some RNA-seq data to the hg19 reference genome sequence. We used blat, which allows 750,000 as the intron size in default, but it was a little slow possibly due to the lack of multithread supporting, 3 days for 1.5 million reads. We have found some other softwares which handle small gap/indel, but such gap/indel are usually too small compared to the intron size. We wonder if there is any other software for alignment over large gaps. Thank you.

[Thomas Doktor]

I would suggest using TopHat: http://tophat.cbcb.umd.edu/

[superligang]

I would suggest using TopHat: http://tophat.cbcb.umd.edu/

Thank you for the suggestions. I thought tophat used bowtie to map reads to existing known transcriptome, didn't know it could also uncover unknown transcript isoforms.

Quoted from tophat manual, "Currently, TopHat does not allow short (fewer than a few nucleotides) insertions and deletions in the alignments it reports."

I get it now. It splits the read, and align them individually first. I will try it.

[sklages]

Maybe, when published, "phaster" will do a good job at it. You may want to ask Phil Green (phrap, cross_match) for "phaster" beta-testing; it is doing a pretty fast gapped alignment. You may also try to use 'cross_match' which is also capable of doing gapped alignments.

http://www.phrap.org/index.html

cheers,
Sven

[superligang]

thank you for the hint

[flxlex]

Or try exonerate: http://www.ebi.ac.uk/~guy/exonerate/

[Jonathan]

Or just build your own library of exon-exon boundary regions and align against it;
We do it, it's fast, and as far as we know accurate;

-Jonathan