QSAR analysis for heterocyclic antifungals.

We perform linear regression analyses on 1202 numerical descriptors that encode the various aspects of the topological, geometrical and electronic molecular structure with the aim of achieving the best QSAR relationship between the antifungal potencies against the Candida albicans strain and the structure of 96 heterocyclic ring derivatives. As a realistic application we employ the model found to predict the biological activity for 60 non-yet measured compounds.