CONICET | Buscador de Institutos y Recursos Humanos

Chagas disease is an endemic Latin American parasitosis associated to poverty and rural populations. It is estimated that about 15 million people in Latin America are infected with Chagas? disease; current chemotherapy for Chagas is effective only in the initial, acute phase of the disease [1]. The progressive increase in the knowledge of the molecular biology of the causal agent (Trypanosoma cruzi) has facilitated the rational development of specific chemotherapies to treat Chagas. The cysteine protease cathepsin L-type named cruzipain stands out among the most promising novel molecular targets to develop innovative antichagasic medications. Cruzipain is the major protease of T. cruzi and it is active in all stages of the parasite life-cycle; thus, cruzipain inhibitors may prove effective in all stages of Chagas [2]. Virtual screening consists in applying computational models to seek drug candidates throughout large virtual chemical repositories in an efficient manner. We present the development of a new model based on molecular topology and capable of predicting whether a given drug candidate is or is not a cruzipain reversible inhibitor. For modeling purposes, 163 cruzipain reversible inhibitors and non-inhibitors were compiled from literature; this dataset was partitioned into representative training and test sets by application of a 2-step clustering analysis based on a hierarchical approach (maximal common substructure) and subsequent k-means clustering. Afterwards, Linear Discriminant Analysis was conducted to derive a binary classificator based on topological descriptors from Dragon Software (Milano Chemometrics). The topological model showed good predictive capability, with an area under the ROC curve of 0.878 for the test set. The obtained topological model may be used to efficiently detect novel antichagasic drugs through large chemical databases.

enviar mensaje