[Bioperl-l] limit the number of blast output per query

[Peter Cock]

On Thu, Jul 7, 2011 at 3:43 AM, Ross KK Leung <ross at cuhk.edu.hk> wrote:
> I know this question should submit to BLAST help but it seems they have
> already been overwhelmed by incoming emails. I wonder any bioperl users
> happen to know how to limit the number of blast output per query. For
> example, for human genome as a database to blast against, a single query can
> generate 10,000+ hits. I have already supplied -b 30 -v 30 flags but
> obviously the blastall from blast2.2.22 does not "obey" my instruction.
>
> The output files generated are usually larger than 100G+ but indeed the
> final ones that I want usually are only of 10M-. Is there any way to help
> save our Earth (Not exaggerated, energy is WASTED in a meaningless
> manner)?

Why are you using such an old version? blastall 2.2.25 is out and may
have fixed this (I expect there is a changelog somewhere), or better yet
at some point you should switch to blast+ rather than continuing with
legacy blast.

Blast+ gives you limits -num_alignments, and -num_descriptions and
-max_target_seqs (from memory the first two only apply to the plain
text output).

Also, perhaps some of the hit property limits like e-value might be
relevant for limiting the number of results.

Peter