[BiO BB] please tell me your opinion about the new BLAST API

Sotiris Salloum ssal at intracom.gr
Thu Dec 15 13:26:12 EST 2005

Dear Christoph, 

Actually Java string processing is very easy, compared to other languages.
Check http://www.sitepoint.com/article/java-regex-api-explained , regarding 
your contribution looks very nice. 

Regarding performance issues with the new 64bit processors, java performance
differences with other languages will be minor, the major thing is
productivity and structured/reusable bioinformatics software.


-----Original Message-----
From: bio_bulletin_board-bounces+ssal=intracom.gr at bioinformatics.org
[mailto:bio_bulletin_board-bounces+ssal=intracom.gr at bioinformatics.org] On
Behalf Of Dr. Christoph Gille
Sent: Thursday, December 15, 2005 5:15 PM
To: bio_bulletin_board at bioinformatics.org
Subject: [BiO BB] please tell me your opinion about the new BLAST API 

Though many things like loops and string processing in Java are still
not as easy and as performant as in other languages Java has to my
opinion gained much attractiveness as a scripting language since
version 1.5.  Java (SUN) is surptisingly as fast as GNU-C/G++ on Intel/AMD
and hence can be used for number crunching (it is really true).

I have just added BLAST functionality to the open source toolbox STRAP
and would like you to have a look at the API.  These are  wrappers for the
EBI server and the local blast programs NCBI-blast and WU-blast.
I could add more in the future.


Please tell me whether it is concise and well structured or whether
there are methods missing.  It is important that I fix problems just
now before other people use the BLAST interface in their projects.

Biojava has been lacking a wrapper for BLAST search sofar but there
are parsers for the XML output in Biojava. Thus STRAP toolbox is
complementary to Biojava and both can work together. Please also send
suggestion for even tighter integrations of STRAP and Biojava.

Life scientists using the GUI for blasting have two advantages over
using a Web mask:

1. Multiple queries: A series of blast jobs for a number of query sequences
   can be started.

2. Cache: Blast results are stored in a HD cache and are computed only once.
   Subsequently, an identical query yields the result immediately.

Please send your comments and suggestions


Bioinformatics.Org general forum  -  BiO_Bulletin_Board at bioinformatics.org

More information about the BBB mailing list