In article <8a6bp5$mpv at darkstar.ucsc.edu>,
Kevin Karplus <karplus at cse.ucsc.edu> wrote:
>>>In article <8a5u30$cdh$1 at niobium.hgmp.mrc.ac.uk>, gwilliam at hgmp.mrc.ac.uk (Gary Williams) writes:
>|> Our sysadmin (Peter Tribble) says:
>|>>|> Getting NCBI blast to actually run reliably seems rather difficult.
>|> Most of the benchmarks I've run have crashed out"
>|>>>This makes me glad I'm running wu-blastp, even if it does still report
>P-values instead of the more useful E-values. Have the NCBI problems
>been with protein searches or with nucleic-acid searches?
We do the vast majority of searches with nucleic acid query sequences.
We see this problem with searches of both nucleic and protein databases
(blastn and blastx searches).
We are considering moving over to using WU-Blast even though it will be
an immense pain changing all our parsing scripts and programs.
Thankfully, we use bioperl.pm a lot now.
Is anyone from NCBI reading this thread?
Would you care to comment?
Gary
