Pharmaceutical targets

Cancer

Our Adaptive Fuzzy Partition method was applied on an anticancer data set, comprised of ca. 1,300 compounds. Those compound were derived from the MDL MDDR database and subdivided into eight mechanisms of actions (antimetabolites, alkylating agents, antracyclines, ...).
The experimental mechanism of action was predicted correctly for about 80% of the test compounds.

Central Nervous System (CNS)

Our fuzzy logic methods were also applied on two CNS data sets, derived from public and commercial data bases (RBI, Tocris, ...), including 400-600 compounds and subdivided into 8 receptor classes (dopaminergic, histaminergic, serotonergic, ...).
We achieved up to 85% of correct class prediction, despite the non specificity of some agonist or antagonist compounds.

Anti-inflammation

We have applied standard molecular modeling procedures (automated docking, CoMFA, 3D QSAR) on large series of indole inhibitors of hnps-PLA2, an enzyme involved in inflammation processes. The prediction models so generated allowed us to build pharmacophores to design new selective compounds with efficient anti-inflammatory properties and negligible side effects.

Note> we have been also successfully involved in the analysis of AIDS and Alzheimer problems.