AI-Assisted Prioritization of Combination Therapies Based on Metabolic Pathway Interactions and Cellular Network Connectivity
Keywords:
drug combination prediction; network medicine; metabolic pathways; graph neural networks; synergy modeling; uncertainty quantificationAbstract
Combination therapy design faces a combinatorial search problem in which the number of possible drug pairs or higher-order regimens vastly exceeds the capacity of experimental testing. Computational prioritization can reduce this burden, yet many existing predictors treat drug synergy as a property of molecular similarity or target overlap and do not explicitly represent the metabolic pathway interactions and cellular network context that shape combination response. This manuscript develops an AI-assisted prioritization framework that integrates metabolic pathway cross-talk with cellular network connectivity to rank candidate combinations. The framework represents drugs, targets, pathways, metabolic reactions, and cell-context nodes in a heterogeneous graph; derives pathway-interaction features from shared reactions and metabolite overlap; computes connectivity scores by network propagation from drug targets; and uses a relational graph neural network to learn context-dependent drug embeddings. A bilinear synergy model combines learned embeddings with pathway-interaction and connectivity features, while Monte Carlo dropout supplies predictive uncertainty for risk-adjusted ranking. The central methodological contribution is a coupled representation in which pathway interactions modify the connectivity features that propagate through the cellular network, rather than being treated as independent covariates. Analytical assessment identifies conditions under which pathway-interaction features are informative, such as convergent metabolic dependencies and compensatory flux, and conditions under which connectivity can mislead, such as network rewiring and cell-type heterogeneity. The framework also specifies baselines, ablation logic, and failure modes, including label noise, batch effects, and distribution shift. By making uncertainty explicit and linking predictions to mechanistic pathway and network quantities, the approach supports triage of combination therapies for experimental follow-up while preserving the distinction between model-derived prioritization and validated biological effect.
References
1. Chou, T. C. (2006). Theoretical basis, experimental design, and computerized simulation of synergism and antagonism in drug combination studies. Pharmacological Reviews, 58(3), 621-681. https://doi.org/10.1124/pr.58.3.10
2. Barabási, A. L., Gulbahce, N., & Loscalzo, J. (2011). Network medicine: a network-based approach to human disease. Nature Reviews Genetics, 12(1), 56-68. https://doi.org/10.1038/nrg2918
3. Hopkins, A. L. (2008). Network pharmacology: the next paradigm in drug discovery. Nature Chemical Biology, 4(11), 682-690. https://doi.org/10.1038/nchembio.118
4. Yıldırım, M. A., Goh, K. I., Cusick, M. E., Barabási, A. L., & Vidal, M. (2007). Drug-target network. Nature Biotechnology, 25(10), 1119-1126. https://doi.org/10.1038/nbt1338
5. Chen, X., Yan, C. C., Zhang, X., & You, Z. H. (2016). Drug-target interaction prediction: databases, web servers and computational models. Briefings in Bioinformatics, 17(4), 696-712. https://doi.org/10.1093/bib/bbv066
6. HerbSyner_Finder: a network community-based model for identifying synergistic combinations from herbal medicines and complex systems. Targetome. 2026;2(2):0-0. doi: 10.48130/targetome-0026-001
7. Ma, T., Xiao, C., Zhou, J., & Wang, F. (2018). Drug similarity integration through attentive multi-view graph auto-encoders. In Proceedings of the 27th International Joint Conference on Artificial Intelligence (IJCAI) (pp. 3477-3483). AAAI Press.
8. Zitnik, M., Agrawal, M., & Leskovec, J. (2018). Modeling polypharmacy side effects with graph convolutional networks. Bioinformatics, 34(13), i457-i466. https://doi.org/10.1093/bioinformatics/bty294
9. Preuer, K., Lewis, R. P., Hochreiter, S., Bender, A., Bulusu, K. C., & Klambauer, G. (2018). DeepSynergy: predicting anti-cancer drug synergy with Deep Learning. Bioinformatics, 34(9), 1538-1546. https://doi.org/10.1093/bioinformatics/btx806
10. Huang, L., Li, F., Sheng, J., Xia, X., Ma, J., Zhan, M., & Wong, S. T. (2014). DrugComboRanker: drug combination discovery based on target network analysis. Bioinformatics, 30(12), i228-i236. https://doi.org/10.1093/bioinformatics/btu278
11. Kipf, T. N., & Welling, M. (2017). Semi-supervised classification with graph convolutional networks. In International Conference on Learning Representations (ICLR).
12. Veličković, P., Cucurull, G., Casanova, A., Romero, A., Lio, P., & Bengio, Y. (2018). Graph attention networks. In International Conference on Learning Representations (ICLR).
13. Schlichtkrull, M., Kipf, T. N., Bloem, P., van den Berg, R., Titov, I., & Welling, M. (2018). Modeling relational data with graph convolutional networks. In European Semantic Web Conference (pp. 593-607). Springer. https://doi.org/10.1007/978-3-319-93417-4_38
14. Gal, Y., & Ghahramani, Z. (2016). Dropout as a Bayesian approximation: Representing model uncertainty in deep learning. In Proceedings of the 33rd International Conference on Machine Learning (pp. 1050-1059). PMLR.
15. Kendall, A., & Gal, Y. (2017). What uncertainties do we need in Bayesian deep learning for computer vision? In Advances in Neural Information Processing Systems (Vol. 30, pp. 5574-5584).
16. O'Boyle, N. M., & Sayle, R. A. (2016). Comparing structural fingerprints using a literature-based similarity benchmark. Journal of Cheminformatics, 8, 36. https://doi.org/10.1186/s13321-016-0148-0
17. Wishart, D. S., Feunang, Y. D., Guo, A. C., Lo, E. J., Marcu, A., Grant, J. R., Sajed, T., Johnson, D., Li, C., Sayeeda, Z., Assempour, N., Iynkkaran, I., Liu, Y., Maciejewski, A., Gale, N., Wilson, A., Chin, L., Cummings, R., Le, D., ... Wilson, M. (2018). DrugBank 5.0: a major update to the DrugBank database for 2018. Nucleic Acids Research, 46(D1), D1074-D1082. https://doi.org/10.1093/nar/gkx1037
18. Kanehisa, M., & Goto, S. (2000). KEGG: Kyoto encyclopedia of genes and genomes. Nucleic Acids Research, 28(1), 27-30. https://doi.org/10.1093/nar/28.1.27
19. Szklarczyk, D., Gable, A. L., Lyon, D., Junge, A., Wyder, S., Huerta-Cepas, J., Simonovic, M., Doncheva, N. T., Morris, J. H., Bork, P., Jensen, L. J., & von Mering, C. (2019). STRING v11: protein-protein association networks with increased coverage. Nucleic Acids Research, 47(D1), D607-D613. https://doi.org/10.1093/nar/gky1131
20. Regev, A., Teichmann, S. A., Lander, E. S., Amit, I., Benoist, C., Birney, E., Bodenmiller, B., Campbell, P., Carninci, P., Clatworthy, M., Clevers, H., Deplancke, B., Dunham, I., Eberwine, J., Eils, R., Enard, W., Farmer, A., Fugger, L., Göttgens, B., ... Human Cell Atlas Meeting Participants. (2017). The human cell atlas. eLife, 6, e27041. https://doi.org/10.7554/eLife.27041
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Bioinformatics Insights and Analytics

This work is licensed under a Creative Commons Attribution 4.0 International License.