[Biophp-dev] AFter Sean's brain stops hurting

S Clark biophp-dev@bioinformatics.org
Sun, 4 May 2003 01:51:10 -0600


Just wanted to give a quick response to this part before I go to bed - more
tomorrow:

On Saturday 03 May 2003 02:33 pm, Serge Gregorio wrote:
[...]
> Also, I've finally thought of an immediate use for your eFetch/
> eUtils set of modules.  I'd like to submit a protein sequence
> (like the one shown above) to NCBI'S BLAST server and get some
> matches, from within a PHP script.  Could you provide some info
> on how your code can be used for this purpose (or how it can be
> extended/modified to do this)?  Or better yet, maybe you could
> do the necessary coding (when your BRAIN STOPS HURTING, that is).

The NCBI Blast query interface is actually different from the EUtils ones, BUT
I have actually already done some preliminary work on it.  If you look
in the "point-n-laugh" subdirectory of the "sean" section, you'll be
able to see what I've got so far.

Basically, I worked out a minimum that was necessary to actually get
the query sent off correctly to NCBI, and to wait for the results, so
other than cleaning up and better organizing it, the only other thing
needed is the parser for the results.

It should be noted that an interface to the NCBI Blast servers will probably
not be very easy to use inside a web page, though just fine for command-line
PHP.  The reason is that the NCBI Blast queries can sometimes take a LONG
time to complete.  Normally it's only 2-3 minutes at most, but I have
had times when queries have taken over an hour to complete (not often, but it
happens.  It's not uncommon to take 10-15 minutes sometimes.)

More in the morning...