HierDTI: A Hierarchical Deep Neural Framework for Drug–Target Interaction Prediction

Mudiyala N S Charishma Reddy; Bomma Venkata Sai Sowrya; Mani Deepak Choudhry

Authors

Mudiyala N S Charishma Reddy Department of Computing Technologies, SRM Institute of Science and Technology, Kattankulathur Campus, Chennai, Tamil Nadu, India Author
Bomma Venkata Sai Sowrya Department of Computing Technologies, SRM Institute of Science and Technology, Kattankulathur Campus, Chennai, Tamil Nadu, India Author
Mani Deepak Choudhry Department of Computing Technologies, SRM Institute of Science and Technology, Kattankulathur Campus, Chennai, Tamil Nadu, India Author

Abstract

This work offers a hierarchical deep learning architecture for the prediction of drug-target protein interactions (DTIs) based on sequence representations of molecular SMILES strings and protein sequences. The new architecture uses multi-level supervision with coarse, intermediate, and final prediction heads. The experiment was conducted under a strict cold-drug split and stratified five-fold experimental validation setup. The model showed an accuracy of 94% and an AUROC of 0.9800 on the cold-drug split. In the five-fold cross-validation setup, the model showed a mean accuracy of 92.96% and a mean AUROC of 0.9774. The results indicate that hierarchical supervision enhances the stability of predictions and generalization performance for trustworthy DTI prediction. Additional improvement in stability-centered architectures may further boost real-world applications in computational drug discovery.

References

[1] M. Jiang, S. Wang, S. Zhang, W. Zhou, Y. Zhang, and Z. Li, “Sequence-based drug-target affinity prediction using weighted graph neural networks,” BMC Genomics, vol. 23, no. 1, p. 449, 2022.

[2] S. Liu, Y. Wang, Y. Deng, L. He, B. Shao, J. Yin, and T. Wang, “Improved drug–target interaction prediction with intermolecular graph transformer,” Briefings in Bioinformatics, vol. 23, no. 5, p. bbac162, 2022.

[3] T. M. Nguyen, T. Nguyen, and T. Tran, “Mitigating cold-start problems in drug-target affinity prediction with interaction knowledge transferring,” Briefings in Bioinformatics, vol. 23, no. 4, p. bbac269, 2022.

[4] J. Bian, X. Zhang, X. Zhang, D. Xu, and G. Wang, “MCANet: shared-weight-based Multihead-CrossAttention network for drug–target interaction prediction,” Briefings in Bioinformatics, vol. 24, no. 2, p. bbad082, 2023.

[5] R. Zhang, Z. Wang, X. Wang, Z. Meng, and W. Cui, “MHTAN-DTI: Metapath-based hierarchical transformer and attention network for drug–target interaction prediction,” Briefings in Bioinformatics, vol. 24, no. 2, p. bbad079, 2023.

[6] X. Zheng, Z. Zhang, X. Zhang, N. Jin, and Y. Zhang, “Ricci-GraphDTA: A graph neural network integrating discrete Ricci curvature for drug–target affinity prediction,” Journal of Molecular Graphics and Modelling, p. 109170, 2025.

[7] H. Zhai, H. Hou, J. Luo, X. Liu, Z. Wu, and J. Wang, “DGDTA: dynamic graph attention network for predicting drug–target binding affinity,” BMC Bioinformatics, vol. 24, no. 1, p. 367, 2023.

[8] M. Gao, D. Zhang, Y. Chen, Y. Zhang, Z. Wang, X. Wang, and J. Song, “GraphormerDTI: a graph transformer-based approach for drug-target interaction prediction,” Computers in Biology and Medicine, vol. 173, p. 108339, 2024.

[9] C. He, Z. Zhao, X. Wang, H. Zheng, L. Duan, and J. Zuo, “Exploring drug-target interaction prediction on cold-start scenarios via meta-learning-based graph transformer,” Methods, vol. 234, pp. 10–20, 2025.

[10] P. M. Shah, H. Zhu, Z. Lu, K. Wang, J. Tang, and M. Li, “DeepDTAGen: a multitask deep learning framework for drug-target affinity prediction and target-aware drugs generation,” Nature Communications, vol. 16, no. 1, p. 5021, 2025.

[11] R. Hu, R. Ge, G. Deng, J. Fan, B. Tang, and C. Wang, “MultiKD-DTA: enhancing drug-target affinity prediction through multiscale feature extraction,” Interdisciplinary Sciences: Computational Life Sciences, vol. 17, no. 3, pp. 555–565, 2025.

[12] M. Yazdani-Jahromi, N. Yousefi, A. Tayebi, E. Kolanthai, C. J. Neal, S. Seal, and O. O. Garibay, “AttentionSiteDTI: an interpretable graph-based model for drug-target interaction prediction using NLP sentence-level relation classification,” Briefings in Bioinformatics, vol. 23, no. 4, p. bbac272, 2022.

[13] W. Yuan, G. Chen, and C. Y. C. Chen, “FusionDTA: attention-based feature polymerizer and knowledge distillation for drug-target binding affinity prediction,” Briefings in Bioinformatics, vol. 23, no. 1, p. bbab506, 2022.

[14] F. Li, Z. Zhang, J. Guan, and S. Zhou, “Effective drug–target interaction prediction with mutual interaction neural network,” Bioinformatics, vol. 38, no. 14, pp. 3582–3589, 2022.

[15] Y. Zhang, Y. Hu, N. Han, A. Yang, X. Liu, and H. Cai, “A survey of drug-target interaction and affinity prediction methods via graph neural networks,” Computers in Biology and Medicine, vol. 163, p. 107136, 2023.

HierDTI: A Hierarchical Deep Neural Framework for Drug–Target Interaction Prediction

Authors

Abstract

References

Published

Issue

Section

License