On 17/07/12 20:32, Galt Barber wrote:
> Hi, Tony!
>
> Reminder: there are many excellent short-read aligners available today.

Hi, Galt.

Yes, of course, and I suggested using BWA or Bowtie when I joined the 
lab. However, we are searching for CRAC chimeras and my colleagues have 
found that BLAST works well for this purpose. They have already tried 
short-read aligners like NovoAlign but, apart from speed, BLAST works 
better. My suggestion is that we used BLAT instead of BLAST to speed 
things up, which is why I've been using it lately. Most recently, we 
have been looking for intronic sequences and that's why I tried to BLAT 
against a 2bit format version of the Ensemble top-level FASTA file.

In case you're interested, I'm working in David Tollervey's lab at the 
University of Edinburgh:

   http://tollervey.bio.ed.ac.uk

[The pipeline I'm working on detects CRAC chimeric hybrid reads]

BLAT has been great for most of the things I've done to date - The 
problem is simply that "blat" segfaults sometimes when searching a 
'large' database (> 4GiB). I understand why people are advising me to 
split the DB into smaller fragments. However, I think it is slightly 
worrying that an otherwise reliable program will sometimes segfault, and 
that this behaviour is almost exactly reproducible by someone else 
running "blat" on different hardware with different query and >4GiB DB.

If I've got time, I'll do some more debugging - It might be a memory 
leak that only shows up when indexing >4GiB DB files. As I said here, 
the same query files run fine against the hg19.2bit DB I downloaded. 
However, using this DB "blat" only uses about 3.9GiB RAM.

Thanks for your helpful suggestion,

   Tony.
_______________________________________________
Genome maillist  -  [email protected]
https://lists.soe.ucsc.edu/mailman/listinfo/genome

Reply via email to