INVESTIGADORES
ALBERCA Lucas NicolÁs
congresos y reuniones científicas
Título:
Cestodes? FABPs: Ligand-based virtual screening and drug repurposing from public libraries.
Autor/es:
RODRIGUEZ, SANTIAGO; ALBERCA, LUCAS NICOLÁS; FALLICO, MAXIMILIANO; PRADA GORI, DENIS; FRANCHINI, GISELA; TALEVI, ALAN
Lugar:
Mendoza
Reunión:
Congreso; XI Congreso de la Sociedad Argentina de Protozoología; 2022
Institución organizadora:
Sociedad Argentina de Protozoología
Resumen:
Echinococcosis and cysticercosis are listed among WHO?s list of Neglected tropical diseases(NTD?s), affecting people in tropical and subtropical areas. Echinococcusgranulosus and Echinococcus multilocularis are the causative agents of cystic and alveolarechinococcosis, respectively, while Taenia solium is the parasitic agent involved in cysticercosis.In general, cestodes present an incomplete lipid metabolism lacking many enzymes involved intheir biosynthesis, so they must obtain these molecules from their hosts. In this sense, FattyAcid Binding Proteins (FABPs) have been proposed as essential for the life cycle of cestodes sincethey are relevant in the traffic and delivery of lipids.Based on the target repurposing paradigm, we have generated ligand-based models to identifydrug repurposing perspectives for the treatment of echinococcocis and cysticercosis. Takinginto consideration the structural similarity between human and cestodes FABPs, we compiledfrom literature 288 compounds whose binding ability against human-aFABP had beenexperimentally established by fluorescence displacement assays. Compounds with IC50 valuesbelow 10 uM were labeled as active and above 20 uM as inactive, thus obtaining a dataset of187 actives and 101 inactives. This dataset was then splitted into a training and a test set usingiRaPCA clustering method.We generated 3000 linear models from our training set using a combination of feature baggingand forward stepwise feature selection; these were validated using Leave Group Out crossvalidationand Y-Randomization and tested against the test set. Moreover, we applied aselective ensemble learning approach to increase the predictive ability and robustness, asverified in retrospective screens performance.The best model ensemble was applied in a virtual screening campaign of several publiccompound libraries, obtaining 106 silico hit compounds. These compounds will be tested invitro against cestodes? FABPs isoforms.