Kleisli
is an integration technology
that is rather suitable in the bioinformatics arena.
We developed a prototype system to more effectively
query protein patents.
This system uses Kleisli to tie together the
following sources to answer queries on protein
patents that are considerably more demanding than simple
free-text search:
(1) the
protein section of the Entrez system at
the National Center for Biotechnology Information;
(2) the
BLAST sequence homology service at the
the National Center for Biotechnology Information;
(3) the WU-BLAST2 sequence homology software from
Washington University;
(4) the
Isite system at the US Patent and Trademark Office ; and
(5) the structural classification of protein database
SCOP
at Cambridge MRC Laboratory of Molecular Biology.
The system is aimed at these three
questions:
Is my protein sequence already patented?
What are the prior arts?
How to broaden my patent claims?
The system provides the following information in response
to these questions:
It can find patented sequences that are similar to yours,
it can try to determine the superfamily of your sequence
and can find patented sequences that are similar to some
sequences in the same superfamily,
it can also find sequences that are similar to unpatented
members in the same superfamily.
An on-line demo of this system can be accessed at
http://adenine.krdl.org.sg:8080/examples/patent.