Hi all,

I am interested in building computermodels about gene/protein interaction networks, using reverse engineering on microarray data. This approach is for example used by the company Gene Network Sciences to find potential targets for drugs (1). They have developed a proprietary inference engine for this goal. In this engine existing knowledge is used to guide (further) inferencing.

Do you know of any public domain initiative to develop such a engine and/or
do you hava any experience with it?

Thank you for your reply in advance!

Cor Lieftink

(1) http://www.gnsbiotech.com/

It depends a lot on how you want to structure your model. It may be possible to use Description Logic reasoners like Pellet http://www.mindswap.org/2003/pellet/ , or forward/backward rule chainers that implement something like the Rete algorithm - like JESS (not free if private) http://herzberg.ca.sandia.gov/jess/ or the one built into JENA http://jena.sourceforge.net/ . For more quantitative style models, look for things that implement SBML like this http://webcell.org/ .