Application of Machine Learning Approaches to Identify New Anticonvulsant Compounds Active in the 6 Hz Seizure Model

GOICOECHEA, S.; SBARAGLINI, M.L.; CHUGURANSKY, S.; MORALES, J. F.; RUIZ, M.E.; TALEVI, A. ; BELLERA, C. L.

São João del-Rei

Workshop; II Latin American Workshop in Computational Neuroscience (LAWCN 2019); 2019

Universidade Federal de São João del-Rei

Abstract. Epilepsy is the second most common chronic brain disorder, affecting65 million people worldwide. According to the NIH’s Epilepsy TherapyScreening Program, evaluation of potential new antiepileptic drug candidatesbegins with assessment of their protective effects in two acute seizure models inmice, the Maximal Electroshock Seizure test and the 6 Hz test. The latter elicitspartial seizures through an electrical stimulus of 44 mA, at which many clinicallyestablished anti-seizure drugs do not suppress seizures. The inclusion of this“high-hurdle” acute seizure assay at the initial stage of the drug identificationphase is intended to increase the probability that agents with improved efficacywill be detected. In this work, we have used machine learning approximations todevelop in silico models capable of identifying novel anticonvulsant drugs withprotective effects in the 6 Hz seizure model. Linear classifiers based on Dragonconformation-independent descriptors were generated through an in-house routine in R environment and validated through standard validation procedures. Theywere later combined through different ensemble learning schemes. The bestensemble comprised the 29 best-performing models combined using the MINoperator. With the objective of finding new drug repurposing opportunities (i.e.identifying second or further therapeutic indications, in our case anticonvulsantactivity, in existing drugs), such model ensemble was applied in a virtual screeningcampaign of DrugBank and Sweetlead databases. 28 approved drugs were identified as potential protective agents in the 6 Hz model. The present study constitutes an example of the use of machine learning approximations to systematicallyguide drug repurposing projects.