[bip] BLAST and FASTA performance

[Erich Schwarz]

On Fri, 18 Jul 2008, Andrew Dalke wrote:

> The rule-of-thumb I learned 10 years ago was to use BLAST
> because while FASTA was a more pure algorithm, the performance
> difference was huge, for little sensitivity gain.  Over the
> years I see most people have chosen BLAST.  Why?

    Because the databases have kept growing, massively, to the point
where it takes an unofficial patch from Finland:

    http://www.csc.fi/english/research/sciences/bioscience/programs/blast/optimized_psiblast

to keep at least some psi-BLAST searches from failing.

    My experience, anyway.  I don't have the numbers, but I have
been doing BlastP et al. routinely from 1992 onward and can attest
that the databases have kept up with the CPUs.


> If it's mostly because of the performance, where is the point
> where it makes sense to switch back?

    When Intel starts marketing 1,000-core CPUs and FASTA is
compiled to effectively use them while running in Ubuntu Linux,
would be my guess.  I.e., maybe by 2013.


--Erich