Prediction And Analysis of Drug-Target Interaction Via Comparative Learning on Heterogeneous Network

Jiarui Mou

doi:10.54097/bj8gt228

Authors

Jiarui Mou

DOI:

https://doi.org/10.54097/bj8gt228

Keywords:

Prediction of drug target interaction; Heterogeneous graph neural network; Heterogeneous biomedical networks; Figure attention; Comparative learning.

Abstract

In this paper, the rapid development of drug target interaction in the field of Computer Science in recent years is analyzed in depth, and Heterogeneous Network-based Contrastive Learning for Drug–Target Interaction (HNCL-DTI) is used as the anchor point for developmental exploration, and its data set limitations and model limitations are expanded, in order to integrate and sort out the technical application and development level in this field, and strive to find better integrated applications that improve and integrate scattered technologies in this field. Drug target interaction prediction is an important research direction of bioinformatics. Traditional methods often ignore the edge characteristics when dealing with heterogeneous biomedical networks, resulting in insufficient prediction accuracy. Some feasible paths to improve the recent methods are discussed. With its limitations and corresponding supplements, it is possible to explore better technical models in the field of drug target interaction, which is expected to help better drug target models in the future.

References

[1]Luo Y, Zhao X, Zhou J. et al. A network integration approach for drug-target interaction prediction and computational drug repositioning from heterogeneous information. Nat Commun 8, 573, 2017. https://doi.org/10.1038/s41467-017-00680-8

[2]Wang X, Ji H, Shi C, Wang B, Ye Y, Cui P, Yu PS. Heterogeneous graph attention network. In Proceedings of the 2019 World Wide Web Conference, 2022–2032, 2019.

[3]Gao M, Zhang D, Chen Y, Li Z, Li X. A unified drug–target interaction prediction framework based on knowledge graph and recommendation system. Brief Bioinform 24(1), bbac578, 2023.

[4]Li Z, Wang Y, Liu H, Zhang X. UniCorn: A unified contrastive learning approach for multi-view molecular representation learning. Adv Neural Inform Process Syst 35, 1–14, 2022.

[5]Li Y, Qiao G, Gao X, Wang G. Supervised graph co-contrastive learning for drug–target interaction prediction. Bioinformatics 38(10), 2847–2854, 2022. https://doi.org/10.1093/bioinformatics/btac164

[6]Hu J, Bewong M, Kvashie S, Zhang W, Nofong VM, Wu G, Feng Z. A heterogeneous network-based contrastive learning approach for predicting drug-target interaction. arXiv, 2024.

[7]Chen Q, Zhang X, Li Y, Wang X, Zhu H, Li J. Accurate and transferable drug–target interaction prediction with DrugLAMP. Bioinformatics 40(12), btae693, 2024. https://doi.org/10.1093/bioinformatics/btae693

[8]Van den Oord A, Li Y, Vinyals O. Representation learning with contrastive predictive coding. Adv Neural Inform Process Syst 32, 6926–6936, 2019.

[9]Chen T, Kornblith S, Norouzi M, Hinton G. A simple framework for contrastive learning of visual representations. In Proceedings of the 37th International Conference on Machine Learning, 1597–1607, 2020.

[10]Bian J, Lu H, Dong G, Wang G. Hierarchical multimodal self-attention-based graph neural network for DTI prediction. Brief Bioinform 25(4), bbae293, 2024.

[11]Chen H, Li X, Zhang Y, Wang J. Leveraging contrastive learning for enhanced node representations in tokenized graph transformers. In Proceedings of the 12th International Conference on Learning Representations, 2024.