SimSpread-Ensemble - Development of an ensemble network-based method for drug discovery

SimSpread is a novel computational method to predict protein–ligand interaction that combines network-based inference with chemical similarity, useful for predicting drug targets, virtual screening, and drug repositioning. This project proposal intents to improve several limitations of SimSpread.
Hypothesis: The combination of predictions obtained from SimSpread that use different similarity cutoffs into a single score using a machine learning (ML) model will increase predictive performance, eliminate empty predictions and eliminate the need to optimize similarity cutoff parameter, resulting in a more straightfoward and user-friendly model to predict drug-target interactions.
Tasks:
• Propose an ensemble predictive model.
• Implement a hyperparameter optimization.
• Compare the predictive performance.
Candidates should have good programming skills.