Perceiver CPI: a nested cross-attention network for compound-protein interaction prediction

Motivation: Compound-protein interaction (CPI) plays an essential role in drug discovery and is performed via expensive molecular docking simulations. Many artificial intelligence-based approaches have been proposed in this regard. Recently, two types of models have accomplished promising results in exploiting molecular information: graph convolutional neural networks that construct a learned molecular representation from a graph structure (atoms and bonds), and neural networks that can be applied to compute on descriptors or fingerprints of molecules. However, the superiority of one method over the other is yet to be determined. Modern studies have endeavored to aggregate information that is extracted from compounds and proteins to form the CPI task. Nonetheless, these approaches have used a simple concatenation to combine them, which cannot fully capture the interaction between such information. Results: We propose the Perceiver CPI network, which adopts a cross-attention mechanism to improve the learning ability of the representation of drug and target interactions and exploits the rich information obtained from extended-connectivity fingerprints to improve the performance. We evaluated Perceiver CPI on three main datasets, Davis, KIBA and Metz, to compare the performance of our proposed model with that of state-of-the-art methods. The proposed method achieved satisfactory performance and exhibited significant improvements over previous approaches in all experiments.

Automated summary

Compound-protein interaction (CPI) is crucial in drug discovery, and there have been AI-based approaches proposed to study it. Two types of models, graph convolutional neural networks and neural networks applied to molecular descriptors or fingerprints, have shown promising results. However, it is still unclear which method is superior. This study presents the Perceiver CPI network, which utilizes a cross-attention mechanism and rich information from extended-connectivity fingerprints to enhance the learning ability and performance. The proposed method outperforms previous approaches in all experiments on three main datasets.