Toward Drug-Target Interaction Prediction via Ensemble Modeling and Transfer Learning

TL;DR

This paper presents EnsembleDLM, an ensemble deep learning approach that predicts drug-target interactions using only sequence data, achieving state-of-the-art results and effective transfer learning across diverse bio-activity types.

Contribution

Introduces EnsembleDLM, a novel ensemble deep learning method that leverages transfer learning for improved drug-target interaction prediction using sequence data.

Findings

Achieves state-of-the-art performance on Davis and KIBA datasets.

Demonstrates effective transfer learning with >0.8 correlation metrics.

Performs well across different bio-activity types and protein classes.

Abstract

Drug-target interaction (DTI) prediction plays a crucial role in drug discovery, and deep learning approaches have achieved state-of-the-art performance in this field. We introduce an ensemble of deep learning models (EnsembleDLM) for DTI prediction. EnsembleDLM only uses the sequence information of chemical compounds and proteins, and it aggregates the predictions from multiple deep neural networks. This approach not only achieves state-of-the-art performance in Davis and KIBA datasets but also reaches cutting-edge performance in the cross-domain applications across different bio-activity types and different protein classes. We also demonstrate that EnsembleDLM achieves a good performance (Pearson correlation coefficient and concordance index > 0.8) in the new domain with approximately 50% transfer learning data, i.e., the training set has twice as much data as the test set.