Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2021 Jul 29;11(8):1119.
doi: 10.3390/biom11081119.

MCN-CPI: Multiscale Convolutional Network for Compound-Protein Interaction Prediction

Shuang Wang  1 , Mingjian Jiang  2 , Shugang Zhang  3 , Xiaofeng Wang  3 , Qing Yuan  3 , Zhiqiang Wei  3 , Zhen Li  4
Affiliations

MCN-CPI: Multiscale Convolutional Network for Compound-Protein Interaction Prediction

Shuang Wang et al. Biomolecules. .

Abstract

In the process of drug discovery, identifying the interaction between the protein and the novel compound plays an important role. With the development of technology, deep learning methods have shown excellent performance in various situations. However, the compound-protein interaction is complicated and the features extracted by most deep models are not comprehensive, which limits the performance to a certain extent. In this paper, we proposed a multiscale convolutional network that extracted the local and global features of the protein and the topological feature of the compound using different types of convolutional networks. The results showed that our model obtained the best performance compared with the existing deep learning methods.

Keywords: compound–protein interaction; convolutional network; deep learning; drug screening.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Figure 1
Figure 1
Overview of the multiscale convolutional network (MCN) for compound–protein interaction prediction.
Figure 2
Figure 2
The descriptor of the binding site. (a)The binding site occupies a small space in the complex of the protein and the molecule; (b) The binding site and the molecule are connected by non-covalent bond; (c) The constructed box of the binding site.
Figure 3
Figure 3
Feature extraction of the binding site based on multi-channel 3D convolutional neural network.
Figure 4
Figure 4
The global feature extraction from the protein based on 1D convolutional neural network.
Figure 5
Figure 5
The prediction of compound–protein interaction.
Figure 6
Figure 6
The performance of the model on different proteins.
Figure 7
Figure 7
The model performance of 5-fold cross-validation for different feature combinations.
Figure 8
Figure 8
The performance of models with different feature combinations on the test dataset.
Figure 9
Figure 9
The model performance of 5-fold cross-validation for different convolutional layers.
See this image and copyright information in PMC

References

    1. Ma D.-L., Chan D.S.-H., Leung C.-H. Drug repositioning by structure-based virtual screening. Chem. Soc. Rev. 2013;42:2130–2141. doi: 10.1039/c2cs35357a. - DOI - PubMed
    1. LeCun Y., Bengio Y., Hinton G. Deep learning. Nature. 2015;521:436. doi: 10.1038/nature14539. - DOI - PubMed
    1. Korshunova M., Ginsburg B., Tropsha A., Isayev O. OpenChem: A Deep Learning Toolkit for Computational Chemistry and Drug Design. J. Chem. Inf. Model. 2021;61:7–13. doi: 10.1021/acs.jcim.0c00971. - DOI - PubMed
    1. Cai C., Wang S., Xu Y., Zhang W., Tang K., Ouyang Q., Lai L., Pei J. Transfer learning for drug discovery. J. Med. Chem. 2020;63:8683–8694. doi: 10.1021/acs.jmedchem.9b02147. - DOI - PubMed
    1. Wang S., Li Z., Zhang S., Jiang M., Wang X., Wei Z. Molecular Property Prediction Based on a Multichannel Substructure Graph. IEEE Access. 2020;8:18601–18614. doi: 10.1109/ACCESS.2020.2968535. - DOI

Publication types

LinkOut - more resources