Genbank FASTA server -> NCBI?

David Mathog mathog at seqvax.caltech.edu
Thu Oct 1 15:24:00 EST 1992


In article <1992Oct1.045436.6970 at nmr-z.mgh.harvard.edu>, cherry at OPAL.MGH.HARVARD.EDU writes...
> 
>I saw nothing in your message about why you do not use the BLAST
>server.
>

Sure didn't mean to give that impression!  We certainly *do* use the BLAST
server and heartily recommend it to all.

> 
>For about the last year BLAST searches have all but replaced FASTA
>searches here at Massachusetts General Hospital. I don't have any
>problem with the FASTA server going away because we didn't use it and
>a great alternative is available (blast at ncbi.nlm.nih.gov).
> 

Hmm.  I suppose if you're only comparing proteins this might be ok, but
since BLAST isn't great at detecting gapped sequences it is not usually the
best choice when running DNA, especially noncoding DNA.

>
>I think you greatly underestimate the amount of horsepower required to
>setup one of these servers. FASTA takes a lot of CPU time.  Opening a
>mail server requires that you can provide and support the service for
>the future.
> 

Not at all.  My point is that if it takes a lot of power to run FASTA/BLAST
then it's probably cheaper to use one really powerful machine at the server
than to set up 50 semipowerful local sites.  It's certainly more efficient
in terms of labor, hardware, cpu cycles, etc. 

>Mike Cherry 
> 
>P.S. I also think your economics of computer purchases is flawed.  Any
>computers or periferals that are purchases are used for much more that
>just sequence searches. Plus in the example of a lab buying a CD-ROM
>drive there are several other diskes that are available and quite
>useful. In some cases not available via any other media: eg. the NCBI
>Entrez CD-ROM, PIR Atlas disk and some hardware vendors routinely
>provide software updates on CD.

I agree that CD-ROMs can be useful, and that the economics changes once
everybody has CD-ROM drives, but at the moment, for all practical purposes,
nobody here has a drive. As far as I know there isn't a single CD-ROM drive
on a Mac either in our department or in the Mac lab at the computer center
(and I've been looking for one ever since I got the NCBI entrez disk). 
There are one or two drives on PCs in the computer center, plus at least 3
on the main campus VAX cluster, and I'm getting one for my VAXstation.  We
can argue various sides of why there are so few drives around, but it's
the way things are.  My guess is that people may soon decide to archive
images via Kodak's photo CDs.  That will put drives in the labs and CD-ROM
based data/ software might take off around here. 

Technology has something to do with this too. If CD-ROM drives cost $200
and had 20 millisecond access times there'd be one on every new machine
coming in the door. 

David Mathog
mathog at seqvax.caltech.edu
manager, sequence analysis facility, biology division, Caltech




More information about the Bio-soft mailing list