[Pipet Devel] BlastXML

Harry Mangalam hjm at cx408397-a.irvn1.occa.home.com
Sun Jan 24 20:54:26 EST 1999


On Sun, 24 Jan 1999, J.W. Bizzaro wrote:

    [most deleted]

> > Many authors of *new and
> > improved* programs, ie FASTA, have included parsible output into their
> > programs.  This makes it easier to connect different analyses.
> 
> Something we must take a look at.  Have you looked into that Harry?

I've been involved in a porting project involving FASTA, but I admit I
missed the bit about making the output more parsable - I'll go back and
check it more carefully - thanks!!  Do others have any other pointers to
packages that have made efforts to make parseable output?

In relation to this is an approach that Lincoln Stein discussed in an article
about using perl for the human genome project which I'll also throw out for
general misinformation: the use of an i/o language called boulderio which
had its beginnings in development of the Whitehead's 'Primer'.  He described
it as a way to pass data thru pipes with each added analyses being able to
tag it with additional info. I'm not suggesting using it as is, but the idea
of being able to add analytical value to a pipes/streams-based dataflow is
vary attractive, especially to a large effort such as a genome initiative or
even pharma.  The article and links to boulderio are at:
http://bio.perl.org/GetStarted/tpj_ls_bio.html
http://stein.cshl.org/software/boulder/

This is a lightweight approach to marking up data so that it can be passed
from app to app.  It is not a very formal approach, but it has been used to
coordinate some very large sequencing efforts.

> 
> > IMHO a good
> > thing.  Most MolbBio packages that I have seen are just a bag of unrelated
> > pieces. meaning you can't run the output of BLAST or FASTA into CLUSTAL
> > without a bit of work.
> 
> I couldn't have said it better myself! :-)  This is what Loci must _not_
> become!  Whatever we use, it must pass this test:  Will it cause a break in
> the Loci continuum?  Will Loci become "a bag of unrelated pieces"?  All Loci
> data (in XML) should be able to be tossed around between the core parts of
> Loci like a basketball at a Harlem Globetrotter's game!

I like that analogy.  

Cheers
Harry



More information about the Pipet-Devel mailing list