Predicting Drug–Target Interactions with Deep-Embedding Learning of Graphs and Sequences
- 18 June 2021
- journal article
- research article
- Published by American Chemical Society (ACS) in The Journal of Physical Chemistry A
- Vol. 125 (25) , 5633-5642
Abstract
Computational approaches for predicting drug–target interactions (DTIs) play an important role in drug discovery since conventional screening experiments are time-consuming and expensive. In this study, we proposed end-to-end representation learning of a graph neural network with an attention mechanism and an attentive bidirectional long short-term memory (BiLSTM) to predict DTIs. For efficient training, we introduced a bidirectional encoder representations from transformers (BERT) pretrained method to extract substructure features from protein sequences and a local breadth-first search (BFS) to learn subgraph information from molecular graphs. Integrating both models, we developed a DTI prediction system. As a result, the proposed method achieved high performances with increases of 2.4% and 9.4% for AUC and recall, respectively, on unbalanced datasets compared with other methods. Extensive experiments showed that our model can relatively screen potential drugs for specific protein. Furthermore, visualizing the attention weights provides biological insight.Keywords
Funding Information
- China Medical University Hospital (DMR-107-067, DMR-108-132, DMR-110-097)
- Guangzhou Municipal Science and Technology Bureau (201803010072)
- Science, Technology and Innovation Commission of Shenzhen Municipality (JCYL 20170818165305521)
This publication has 32 references indexed in Scilit:
- Improving compound–protein interaction prediction by building up highly credible negative samplesBioinformatics, 2015
- The RCSB Protein Data Bank: views of structural biology for basic and applied research and educationNucleic Acids Research, 2014
- Lessons Learned in Empirical Scoring with smina from the CSAR 2011 Benchmarking ExerciseJournal of Chemical Information and Modeling, 2013
- Directory of Useful Decoys, Enhanced (DUD-E): Better Ligands and Decoys for Better BenchmarkingJournal of Medicinal Chemistry, 2012
- Systematic Generation of Chemical Structures for Rational Drug Design Based on QSAR ModelsCurrent Computer-Aided Drug Design, 2011
- AutoDock Vina: Improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreadingJournal of Computational Chemistry, 2009
- GROMACS 4: Algorithms for Highly Efficient, Load-Balanced, and Scalable Molecular SimulationJournal of Chemical Theory and Computation, 2008
- Molecular Docking2008
- Support-vector networksMachine Learning, 1995
- An Introduction to Kernel and Nearest-Neighbor Nonparametric RegressionThe American Statistician, 1992