Article Details

Abstract

Drug response prediction is an important task in precision medicine and relies on effective representations of drug structures and genomic features of cell lines. To address the limitations of SMILES-based methods in preserving complete structural information and of conventional GCNs that ignore chemical bond information, this study proposes BondGCN-DRP. The model represents each drug as a molecular graph with node and edge features, incorporates bond type, bond order, aromaticity, and ring information, and uses a one-dimensional convolutional network to learn cell-line features, followed by feature fusion for drug response prediction. Experiments on the GDSC dataset show that BondGCN-DRP outperforms both tCNNS and standard GCN models in terms of RMSE and Pearson correlation coefficient, while saliency map analysis further demonstrates its interpretability.

Keywords

Deep Learning; Graph Neural Networks; edge feature; drug response prediction; molecular graph modeling; interpretability

References

1. [1] Wang Cong Research on Drug Reaction Prediction Based on Deep Learning [D]. Harbin: Harbin Institute of Technology, 2023
2. [2] Yang W，Soares J，Greninger P，et al. Genomics of Drug Sensitivity in Cancer (GDSC): A Resource for Therapeutic Biomarker Discovery in Cancer Cells [J].Nucleic Acids Research，2013，41（D1）：D955-D961.
3. [3] Barretina J，Caponigro G，Stransky N，et al. Addendum: The Cancer Cell Line Encyclopedia Enables Predictive Modelling of Anticancer Drug Sensitivity [J/OL].Nature，2012，492（7428）：290.https://doi.org/10.1038/nature11735.
4. [4] Deng Dexun Research on Knowledge Guided Deep Learning Model for Predicting Drug Reactions [D]. Hengyang: Nanhua University, 2024
5. [5] Liao Xin Research on Tumor Drug Response Prediction Based on Transformer and Graph Attention Network [D]. Qingdao: Qingdao University of Science and Technology, 2024
6. [6] Wu Juhua, Li Junfeng, Tao Lei Prediction of Adverse Drug Reactions Based on Knowledge Graph Embedding and Deep Learning [J]. Journal of Guangdong University of Technology, 2024, 41 (1): 19-26, 40
7. [7] Liu P，Li H，Li S，et al. Improving Prediction of Phenotypic Drug Response on Cancer Cell Lines Using Deep Convolutional Network [J].BMC Bioinformatics，2019，20（1）：1-14.
8. [8] Nguyen T N，Le H，Quang P T，et al. GraphDTA: Predicting Drug-Target Binding Affinity With Graph Neural Networks [J].Bioinformatics，2021，37（8）：1140-1147.
9. [9] Simonyan K，Vedaldi A，Zisserman A. Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps [PP/OL].（2014-04-19）.https://arxiv.org/abs/1312.6034.
10. [10] O'Boyle N M. Towards a Universal SMILES Representation: A Standard Method to Generate Canonical SMILES Based on the InChI [J/OL].Journal of Cheminformatics，2012，4（1）：22.https://doi.org/10.1186/1758-2946-4-22.
11. [11] Kipf T N，Welling M. Semi-Supervised Classification With Graph Convolutional Networks [PP/OL].（2017-02-22）.https://arxiv.org/abs/1609.02907.
12. [12] Zhang T，Liu B，Niu D，et al. Multiresolution Graph Attention Networks for Relevance Matching [C]//Proceedings of the 27th ACM International Conference on Information and Proceedings of the 27th ACM International Conference on Information and Knowledge Management，2018：933-942.
13. [13] Niepert M，Ahmed M，Kutzkov K. Learning Convolutional Neural Networks for Graphs [PP/OL].（2016-06-08）.https://arxiv.org/abs/1605.05273.
14. [14] Klein T，Nabi M. Attention Is (Not) All You Need for Commonsense Reasoning [PP/OL].（2019-05-31）.https://arxiv.org/abs/1905.13497.

About the authors

Ouyang Fei (1996-), male, Han nationality, from Chengdu, Sichuan, assistant engineer, master's student, research direction: deep learning, computer vision.