The effect of Ebola virus disease (EVD) is fatal and devastating, necessitating several efforts to identify potent biotherapeutic molecules. This review seeks to provide perspectives on complementing existing work on Ebola virus (EBOV) by discussing the role of machine learning (ML) techniques in the prediction of small molecule inhibitors of EBOV. Different ML algorithms have been used to predict anti-EBOV compounds, including Bayesian, support vector machine, and random forest algorithms, which present strong models with credible outcomes. The use of deep learning models for predicting anti-EBOV molecules is underutilized; therefore, we discuss how such models could be leveraged to develop fast, efficient, robust, and novel algorithms to aid in the discovery of anti-EBOV drugs. We further discuss the deep neural network as a plausible ML algorithm for predicting anti-EBOV compounds. We also summarize the plethora of data sources necessary for ML predictions in the form of systematic and comprehensive high-dimensional data. With ongoing efforts to eradicate EVD, the application of artificial intelligence-based ML to EBOV drug discovery research can promote data-driven decision making and may help to reduce the high attrition rates of compounds in the drug development pipeline.
Keywords: Ebola virus; artificial intelligence; big data; classifiers; deep learning; drug discovery; machine learning.