TY - JOUR
T1 - CAT-CPI
T2 - Combining CNN and transformer to learn compound image features for predicting compound-protein interactions
AU - Qian, Ying
AU - Wu, Jian
AU - Zhang, Qian
N1 - Publisher Copyright:
Copyright © 2022 Qian, Wu and Zhang.
PY - 2022/9/15
Y1 - 2022/9/15
N2 - Compound-protein interaction (CPI) prediction is a foundational task for drug discovery, which process is time-consuming and costly. The effectiveness of CPI prediction can be greatly improved using deep learning methods to accelerate drug development. Large number of recent research results in the field of computer vision, especially in deep learning, have proved that the position, geometry, spatial structure and other features of objects in an image can be well characterized. We propose a novel molecular image-based model named CAT-CPI (combining CNN and transformer to predict CPI) for CPI task. We use Convolution Neural Network (CNN) to learn local features of molecular images and then use transformer encoder to capture the semantic relationships of these features. To extract protein sequence feature, we propose to use a k-gram based method and obtain the semantic relationships of sub-sequences by transformer encoder. In addition, we build a Feature Relearning (FR) module to learn interaction features of compounds and proteins. We evaluated CAT-CPI on three benchmark datasets—Human, Celegans, and Davis—and the experimental results demonstrate that CAT-CPI presents competitive performance against state-of-the-art predictors. In addition, we carry out Drug-Drug Interaction (DDI) experiments to verify the strong potential of the methods based on molecular images and FR module.
AB - Compound-protein interaction (CPI) prediction is a foundational task for drug discovery, which process is time-consuming and costly. The effectiveness of CPI prediction can be greatly improved using deep learning methods to accelerate drug development. Large number of recent research results in the field of computer vision, especially in deep learning, have proved that the position, geometry, spatial structure and other features of objects in an image can be well characterized. We propose a novel molecular image-based model named CAT-CPI (combining CNN and transformer to predict CPI) for CPI task. We use Convolution Neural Network (CNN) to learn local features of molecular images and then use transformer encoder to capture the semantic relationships of these features. To extract protein sequence feature, we propose to use a k-gram based method and obtain the semantic relationships of sub-sequences by transformer encoder. In addition, we build a Feature Relearning (FR) module to learn interaction features of compounds and proteins. We evaluated CAT-CPI on three benchmark datasets—Human, Celegans, and Davis—and the experimental results demonstrate that CAT-CPI presents competitive performance against state-of-the-art predictors. In addition, we carry out Drug-Drug Interaction (DDI) experiments to verify the strong potential of the methods based on molecular images and FR module.
KW - compound-protein interaction
KW - deep learning
KW - drug-drug interaction
KW - molecular image
KW - transformer encoder
UR - https://www.scopus.com/pages/publications/85139221812
U2 - 10.3389/fmolb.2022.963912
DO - 10.3389/fmolb.2022.963912
M3 - 文章
AN - SCOPUS:85139221812
SN - 2296-889X
VL - 9
JO - Frontiers in Molecular Biosciences
JF - Frontiers in Molecular Biosciences
M1 - 963912
ER -