Efficient prediction of drug–drug interaction using deep learning models

• Published in: IET systems biology Vol. 14; no. 4; pp. 211 - 216
• Main Authors: Kumar Shukla, Prashant, Kumar Shukla, Piyush, Sharma, Poonam, Rawat, Paresh, Samar, Jashwant, Moriwal, Rahul, Kaur, Manjit
• Format: Journal Article
• Language: English
• Published: England The Institution of Engineering and Technology 01.08.2020
• Subjects: cancer; cancer treatment; Computational Biology; convolutional mixture density recurrent neural network; convolutional neural nets; Deep Learning; drug delivery systems; Drug Interactions; drug synergy; drugs; drug–drug interaction prediction; drug–drug interaction score; Humans; learning (artificial intelligence); machine learning; Neural Networks, Computer; recurrent neural nets; Research Article; drug delivery systems; cancer treatment; drugs; recurrent neural nets; drug–drug interaction prediction; convolutional mixture density recurrent neural network; drug–drug interaction score; machine learning; deep learning; drug synergy; cancer; convolutional neural nets; learning (artificial intelligence)
• ISSN: 1751-8849; 1751-8857; 1751-8857
• DOI: 10.1049/iet-syb.2019.0116

A drug–drug interaction or drug synergy is extensively utilised for cancer treatment. However, prediction of drug–drug interaction is defined as an ill-posed problem, because manual testing is only implementable on small group of drugs. Predicting the drug–drug interaction score has been a popular research topic recently. Recently many machine learning models have proposed in the literature to predict the drug–drug interaction score efficiently. However, these models suffer from the over-fitting issue. Therefore, these models are not so-effective for predicting the drug–drug interaction score. In this work, an integrated convolutional mixture density recurrent neural network is proposed and implemented. The proposed model integrates convolutional neural networks, recurrent neural networks and mixture density networks. Extensive comparative analysis reveals that the proposed model significantly outperforms the competitive models.