FastQ processing

[Ed Hollox]

Hi all

We may be splashing out on an IonTorrent here in Leicester. How does everyone cope with the FASTQ files, and what flavour FASTQ format are they in? I'm considering starting a pipeline using Galaxy and LastZ, but are there any other remapping softwares that do well with longer reads (i.e. not the short 35bp reads that Bowtie etc are optimised for?)

Ed

[lh3]

bwa-sw/ssaha2/blat

I tried lastz before, but could not get it running properly, so I cannot comment how bwa-sw/ssaha2/blat is compared to lastz. 1kg use ssaha2.

[colindaven]

Sanger FastQ is the default. See the Wikipedia page to clear up any uncertainty about quality encoding if you're unfamiliar with it.

Novoalign should also do quite well with reads of that length, as well as perhaps Stampy. Not sure how Galaxy will work out performancewise if you are generating data very regularly, it might be better to have a multithreaded aligner working locally - especially if they are able to ramp up output as much as they claim.

[Ed Hollox]

Thanks both. Colindaven - we will be installing galaxy locally on a research cluster.

[IonTorrent]

TMAP is a freely available mapping software for Ion Torrent data. Released under the GPLv2, the source code is available at the Torrent Dev section of ioncommunity at http://ioncommunity.iontorrent.com.

At Torrent Dev, we have details on FASTQ formats produced in Torrent Suite software, in case anyone has questions on those formats.

We understand that the question of what aligners work best with Ion Torrent data. We had to answer that question ourselves, and we developed TMAP as a reasonable answer to this question. We fully anticipate other aligners to support Ion Torrent data as the platform matures. We are excited to work with both users and developers to improve TMAP for Ion Torrent data, as well as answer questions to help other alignment tools support semiconductor sequencing data.

[Tectona]

T-MAP details

Does TMAP align sequence reads without a reference map, or does it need a reference? I'd like to sequence reduced representation libraries (like RAD-TAG) from multiple individuals belonging to species for which a reference sequence is not available. Would T-MAP be able to align these sequence tags just on their own.. and I could then identify SNPs from them directly?

[nilshomer]

Does TMAP align sequence reads without a reference map, or does it need a reference? I'd like to sequence reduced representation libraries (like RAD-TAG) from multiple individuals belonging to species for which a reference sequence is not available. Would T-MAP be able to align these sequence tags just on their own.. and I could then identify SNPs from them directly?

No, it compares the reads to a reference (aligner), whereas I think you are asking for an assembler.

[KevinLam]

Is there a direct link to download Tmap for a test drive?

Had a hard time finding the torrent dev link
and then i had to register to enter.

here's direct link for benefit of others

http://lifetech-it.hosted.jivesoftware.com/community/torrent_dev

anyway, I am interested in the ram requirements and also just an offshoot, is there any reason why BWA or bowtie won't work with the Ion Torrent FastQ format?