Drug Synergy Prediction Using Dynamic Mutation Based Differential Evolution

Abstract

Purpose: In cancer therapies, drug combinations have shown significant accuracy and minimal side effects than the single drug administration. Therefore, drug synergy has drawn great interest from pharmaceutical companies and researchers. Unfortunately, the prediction of drug synergy score was carried out based on the small group of drugs.

Methods: Due to the advancement in high-throughput screening (HTS), the size of drug synergy datasets has grown enormously in recent years. Hence, machine learning models have been utilized to predict the drug synergy score. However, the majority of these machine learning models suffer from over-fitting and hyperparameters tuning issues.

Results: A novel deep bidirectional mixture density network (BMDN) model is proposed. A dynamic mutationbased multi-objective differential evolution is used to optimize the hyper-parameters of BMDN. Extensive is conducted on the NCI-ALMANAC drug synergy dataset that consists of 2,90,000 synergy determinations.

Conclusions: Experimental results reveal that BMDN outperforms the existing drug synergy models in terms of various performance metrics.

Keywords: BMDN; Drug synergy; HTS; deep learning; machine learning; neural networks.