Trypanosoma brucei rhodesiense causes the most virulent form of sleeping sickness in East Africa. Without treatment this disease has a fatal course. Currently, there are only five drugs in use and only one of them (melarsoprol) is effective in the final stage of the disease. To find new drugs with anti-trypanosomal activity, it has been developed a mathematical model based on molecular topology aimed at finding novel active compounds from a group of bicyclic derivatives of 4-methylpiperazine. Using linear discriminant analysis it was found a model that could correctly classify the activity of 93% of 32 compounds studied. In order to predict the antiparasitic activity, a multilinear regression analysis was carried out so that it was able to explain 86% of the variance. The validation of this model was made by cross-validation and a test of randomness. Finally, the selected model was applied to the search for new bicyclic compounds potentially active.

Keywords: Trypanosoma brucei rhodesiense, QSAR Analysis, molecular topology.