Spliced alignment with BWA

[telos]

Does anyone know how to do spliced alignment (i.e. with N's in the CIGAR string) on colour-space data with BWA? I've had no luck even with -e 100000 -O 2 -E 0: I don't get a single spliced alignment with those options. I've also tried a trial version of NovoalignCS but it crashes and haven't heard back from their developers. TopHat doesn't do colour-space. So basically I'm struggling to do intron-skipping gapped alignment on colour-space data at the moment... Any help would be appreciated.

Thanks.

[KevinLam]

Can you confirm about tophat?

there's a webinar about solid data and tophat here

https://learn.appliedbiosystems.com/solid/

SOLiD™ RNA-Seq Data Analysis featuring TopHat and Cufflinks (1 hour)

If you do not get an answer now perhaps you can try the webinar.

I have to grapple with this soon as well. Do keep me posted on your findings

another resource you can check out is here

http://rna-seqblog.com/tag/transcriptome-sequencing/

Cheers
Kevin

[kopi-o]

TopHat was supposed to add color space support during spring 2010, but there is no update to that effect on the homepage. It would be very nice if the support had in fact been added, as the SOLiD webinar title suggests.

I have had similar troubles with BWA and spliced alignments ("chimeras") using color space reads - I did get a few chimeric alignments but far less than expected (just a handful in total).

You might want to try the GEM library, which has a split read mapper which should work on color space reads.

[krobison]

Anyone actually used GEM for colorspace data? The documentation is, shall we say, sparse. I'm not clear how to set up the index

[krobison]

BTW, the webinar apparently isn't until the 29th, so perhaps TopHat support for colorspace is imminent.

[mmartin]

There is an announcement on the TopHat homepage: It seems color space support has very recently been added.

[KevinLam]

Yups!

TopHat 1.1.0 release 10/03/2010
This release of TopHat includes some major enhancements:
TopHat now supports Colorspace reads from Life Technologies' SOLiD sequencer, thanks to the efforts of Daehwan Kim and Geo Pertea. Note that you will need to download a Colorspace Bowtie index to use TopHat with SOLiD reads. Colorspace Bowtie indices typically have "_c" in their name to distinguish them from nucleotide space indices.

[telos]

This is good news but I've had no luck actually using it.

First of all, it throws an error if the .csfasta and .qual files have #-comments. I had to strip them out.

Second, prep_reads complains:
"Saw a space but expected an ASCII-encoded quality value.
Are quality values formatted as integers? If so, try --integer-quals.
terminate called after throwing an instance of 'int'"

Yet --integer-quals was switched on (since I was using -C).