AttentionDTA: drug-target binding affinity prediction by sequence-based deep learning with attention mechanism.

The prediction of drug-target affinities (DTAs) is substantial in drug development. Recently, deep learning has made good progress in the prediction of DTAs. Although relatively effective, due to the black-box nature of deep learning, these models are less biologically interpretable. In this study, we proposed a deep learning-based model, named AttentionDTA, with attention mechanism. The novelty of our work is to use attention mechanism to focus on key subsequences which are important in drug and protein sequences when predicting its affinity. We use two separate one-dimensional Convolution Neural Networks to extract the semantic information of drug's SMILES string and protein's amino acid sequence. Furthermore, four different attention mechanisms are developed and embedded to our model to explore the relationship between drug features and protein features. We conduct extensive experiments to demonstrate that AttentionDTA can effectively extract protein features related to drug information and drug features related to protein information to better predict drug target affinities. By visualizing the attention weight in the model, we found that even if the information of the binding site was never input during the inference process, AttentionDTA can still effectively enhance the role of the protein feature at the target site.