[Bioperl-l] Fwd: question about Bio::Tools::Run::StandAloneBlast

[Chris Fields]

On Apr 14, 2010, at 10:43 AM, Robert Bradbury wrote:

> On Wed, Apr 14, 2010 at 9:16 AM, Chris Fields <cjfields at illinois.edu> wrote:
> 
>> I'm assuming that we will slowly be deprecating use of those modules as
>> users migrate over to BLAST+.  I'm not sure, but I don't think NCBI will
>> support legacy BLAST that much longer.
>> 
>> 
> By "legacy BLAST" do you refer to blast submission requests from the
> "public"?

No.  The changeover there should be relatively transparent.


> I ask because I use BLAST @ NCBI a fair amount and have scripts that use it.
> [Though I would be very happy for a "Blast Gateway" [Google SoC
> opportunity???] that would run queries against NCBI, JGI, Ensembl & perhaps
> the various Yeast/Insect/Nematode databases].  I state this because of the
> genome scale efforts that I'm aware of Broad seems to be feeding the data to
> Ensembl first while a lot of the data from JGI seems to take a long time to
> make it to NCBI.]
> 
> I do realize that there is an some kind of an exponential growth problem
> with the number of genomes sequenced and blast scans against the entire set
> of genomes but I would hope that computing capacity is scaling with this.
> After all the number of species is finite.

How so?  Do you have a reference for that?  There is this ongoing phenomenon I keep hearing of, called 'evolution'... (or, more specifically, 'speciation'...).

> But it would that if "free" resources for running blast scans are limited
> that interfaces are going to have to move in the direction of a combined
> "public"+"private" ("remote"+"local") methodology.
> 
> R.


Reality is, there will always be a public interface for some users, but there will similarly be local databases or power users who will require local BLAST capability for whatever reason (private data, large-scale analyses, etc).  If there were a demand for such a combined 'public:private' or similar service it certainly hasn't manifested itself yet.

chris