GanDTI: A multi-task neural network for drug-target interaction prediction

[Display omitted] •We present a simple and multi-purpose model, GanDTI, that could both predict the binding affinity and classify the interaction.•The model employs a residual graph neural network to process the compound fingerprint data and forms a vector that could project product-based attention on the protein sequence to determine the binding importance on the sequence. Then the two parts of the data are concatenated for MLP processing.•We evaluate the method on three public benchmark datasets and we find that the model outperforms the state-of-the-art deep learning methods in both tasks after comparisons using various metrics. Drug discovery processes require drug-target interaction (DTI) prediction by virtual screenings with high accuracy. Compared with traditional methods, the deep learning method requires less time and domain expertise, while achieving higher accuracy. However, there is still room for improvement for higher performance with simplified structures. Meanwhile, this field is calling for multi-task models to solve different tasks. Here we report the GanDTI, an end-to-end deep learning model for both interaction classification and binding affinity prediction tasks. This model employs the compound graph and protein sequence data. It only consists of a graph neural network, an attention module and a multiple-layer perceptron, yet outperforms the state-of-the art methods to predict binding affinity and interaction classification on the DUD-E, human, and bindingDB benchmark datasets. This demonstrates our refined model is highly effective and efficient for DTI prediction and provides a new strategy for performance improvement.