ENHANCED GNN WITH CAUSAL PROXIMITY VECTORS: BRIDGING CAUSALITY AND PROXIMITY IN GRAPH NEURAL NETWORKS
DOI:
https://doi.org/10.7494/csci.2026.27.1.6899Abstract
A knowledge graph is a structured representation of entities and their relationships, often used in biomedical domains to model complex interactions. Graph Neural Networks (GNNs), which utilize these graphs, are effective for predicting interactions missing in the knowledge graph. However, GNN lacks the ability to incorporate causal reasoning, which is crucial in biomedical applications. Additionally, they limit their ability to generalize to unseen data.
In oncology, where treatment regimens are intricate and patient responses are highly variable, predicting Adverse Drug Reactions (ADRs) is particularly difficult. Existing models fail to capture the indirect, high-granularity information needed for accurate ADR prediction. To address these challenges, we propose the Causality and Proximity-based Relational Multihead Attention Model (CPRMAM). This model leverages a knowledge graph of ADR-related cancer case studies and introduces a causal proximity vector to prioritize relevant relationships. By employing an inductive GNN approach, CPRMAM generalizes to unseen data, improving ADR prediction.
Downloads
References
[1] Ahmed N.K., Rossi R., Lee J.B., Willke T.L., Zhou R., Kong X., Eldardiry H.:
Learning Role-based Graph Embeddings, arXiv preprint arXiv:180202896, 2018.
https://arxiv.org/abs/1802.02896.
[2] Atias N., Sharan R.: An Algorithmic Framework for Predicting Side Effects
of Drugs, Journal of Computational Biology, vol. 18(3), pp. 207–218, 2011.
doi: 10.1089/cmb.2010.0255.
[3] Bate A., Lindquist M., Edwards I., Olsson S., Orre R., Lansner A., DeFreitas:
A Bayesian Neural Network Method for Adverse Drug Reaction Signal Gener-
ation, European Journal of Clinical Pharmacology, vol. 54, pp. 315–321, 1998.
doi: 10.1007/s002280050466.
[4] Caufield J.H.: MACCROBAT, Dataset, 2019. doi: 10.6084 /
m9.figshare.9764942.v2.
[5] Craswell N.: Mean Reciprocal Rank. In: L. Liu, M.T. ¨Ozsu (eds.), Encyclopedia
of Database Systems, Springer, Boston, MA, 2009. doi: 10.1007/978-0-387-39940-
9 488.
[6] Dev S., Sharan A.: Annotated and Normalized Causal Relation Extraction Cor-
pus for Improving Health Informatics. In: Proceedings of the 20th International
Conference on Natural Language Processing (ICON), pp. 417–422, NLP Associ-
ation of India (NLPAI), 2023.
[7] Dev S., Sharan A.: Automatic Construction of Named Entity Corpus for Adverse
Drug Reaction Prediction. In: A. Bhattacharya, S. Dutta, P. Dutta, V. Piuri
(eds.), Innovations in Data Analytics. ICIDA 2022, Advances in Intelligent Sys-
tems and Computing, vol. 1442, Springer, Singapore, 2023. doi: 10.1007/978-
981-99-0550-8 20.
[8] Dey S., Luo H., Fokoue A., Hu J., Zhang P.: Predicting Adverse Drug Re-
actions through Interpretable Deep Learning Framework, BMC Bioinformatics,
vol. 19(Suppl 21), p. 476, 2018. doi: 10.1186/s12859-018-2544-0.
[9] Gao Y., Ji S., Zhang T., Tiwari P., Marttinen P.: Contextualized Graph Em-
beddings for Adverse Drug Event Detection. In: M. Amini, S. Canu, A. Fischer,
T. Guns, P. Kralj Novak, G. Tsoumakas (eds.), Machine Learning and Knowledge
Discovery in Databases. ECML PKDD 2022, Lecture Notes in Computer Science,
vol. 13714, Springer, Cham, 2023. doi: 10.1007/978-3-031-26390-3 35.
[10] Gao Y., Zhang X., Sun Z., Chandak P., Bu J., Wang H.: Precision Adverse
Drug Reactions Prediction with Heterogeneous Graph Neural Network, Advanced
Science, vol. 12(4), p. e2404671, 2024. doi: 10.1002/advs.202404671. Epub ahead
of print.
[11] Khan M., Mello G., Habib L., Engelstad P., Yazidi A.: HITS-based Propaga-
tion Paradigm for Graph Neural Networks, ACM Transactions on Knowledge
Discovery from Data, vol. 18(4), 2024. doi: 10.1145/3638779.
[12] Kwak H., Lee M., Yoon S., Chang J., Park S., Jung K.: Drug-Disease Graph:
Predicting Adverse Drug Reaction Signals via Graph Neural Network with Clini-
cal Data. In: Advances in Knowledge Discovery and Data Mining, Lecture Notes
in Computer Science, vol. 12085, pp. 633–644, 2020. doi: 10.1007/978-3-030-
47436-2 48.
[13] Lerch M., Nowicki P., Manlik K., Wirsching G.: Statistical Signal Detection as
a Routine Pharmacovigilance Practice: Effects of Periodicity and Resignalling
Criteria on Quality and Workload, Drug Safety, vol. 38, pp. 1219–1231, 2015.
doi: 10.1007/s40264-015-0345-1.
[14] Lin J., Kuang Q., Li Y., Zhang Y., Sun J., Ding Z., Li M.: Prediction of Adverse
Drug Reactions by a Network Based External Link Prediction Method, Analytical
Chemistry, vol. 5(21), p. 6120, 2013. doi: 10.1039/c3ay41290c
[15] Liu B.M., Gao Y.L., Li F., Zheng C.H., Liu J.X.: SLGCN: Structure-
enhanced Line Graph Convolutional Network for Predicting Drug–Disease As-
sociations, Knowledge-Based Systems, vol. 283, p. 111187, 2023. doi: 10.1016/
j.knosys.2023.111187.
[16] Liu R., AbdulHameed M., Kumar K., Yu X., Wallqvist A., Reifman J.: Data-
driven Prediction of Adverse Drug Reactions Induced by Drug-Drug Interactions,
BMC Pharmacology and Toxicology, vol. 18(1), p. 44, 2017. doi: 10.1186/s40360-
017-0153-6.
[17] Manzi S., Reis B.: Predicting Adverse Drug Events Using Pharmacological Net-
work Models, Science Translational Medicine, vol. 3(114), p. 114ra127, 2011.
doi: 10.1126/scitranslmed.3002073.
[18] Miao R., Yang Y., Ma Y., Juan X., Xue H., Tang J., Wang X.: Negative Samples
Selecting Strategy for Graph Contrastive Learning, Information Sciences, vol.
613, pp. 667–681, 2022. doi: 10.1016/j.ins.2022.09.024.
[19] Monaco L., Melis M., Biagi C., et al.: Signal Detection Activity on EudraVig-
ilance Data: Analysis of the Procedure and Findings from an Italian Regional
Centre for Pharmacovigilance, Expert Opinion on Drug Safety, vol. 16, pp. 271–
275, 2017. doi: 10.1080/14740338.2017.1284200.
[20] Pauwels E., Stoven V., Yamanishi Y.: Predicting Drug Side-Effect Profiles: A
Chemical Fragment-Based Approach, BMC Bioinformatics, vol. 12(1), p. 169,
2011. doi: 10.1186/1471-2105-12-169.
[21] Schuemie M.J., Ryan P.B., DuMouchel W., Suchard M.A., Madigan D.: Using
Electronic Health Care Records for Drug Safety Signal Detection: A Compara-
tive Evaluation of Statistical Methods, Medical Care, vol. 50, pp. 890–897, 2012.
doi: 10.1097/mlr.0b013e31825f63bf.
[22] Wang S.V., Maro J.C., Baro E., et al.: Data Mining for Adverse Drug Events with
a Propensity Score-Matched Tree-Based Scan Statistic, Epidemiology, vol. 29,
pp. 895–903, 2018. doi: 10.1097/ede.0000000000000907.
[23] Wishart D.S., Knox C., Guo A.C., Cheng D., Shrivastava S., Tzur D., Gautam
B., Hassanali M.: DrugBank: A Knowledgebase for Drugs, Drug Actions and
Drug Targets, Nucleic Acids Research, vol. 36(Database issue), pp. D901–D906,
2008. doi: 10.1093/nar/gkm958.
[24] Xiang Y.P., Liu K., Cheng X.Y., Cheng C., Gong F., Pan J.B., Ji Z.L.: Rapid As-
sessment of Adverse Drug Reactions by Statistical Solution of Gene Association
Network, IEEE/ACM Transactions on Computational Biology and Bioinformat-
ics, vol. 12(4), pp. 844–850, 2015. doi: 10.1109/tcbb.2014.2338292.
[25] Xiao C., Zhang P., Chaowalitwongse W.A., Hu J., Wang F.: Adverse Drug Reac-
tion Prediction with Symbolic Latent Dirichlet Allocation. In: Proceedings of the
Thirty-First AAAI Conference on Artificial Intelligence (AAAI’17), pp. 1590–
1596, AAAI Press, 2017. doi: 10.1609/aaai.v31i1.10717
[26] Xue R., Liao J., Shao X., Han K., Long J., Shao L., Ai N., Fan X.: Prediction of
Adverse Drug Reactions by Combining Biomedical Tripartite Network and Graph
Representation Model, Chemical Research in Toxicology, vol. 33(1), pp. 202–210,
2020. doi: 10.1021/acs.chemrestox.9b00238.
[27] Yildirim P., Majnari´c L., Ekmekci O., Holzinger A.: Knowledge Discovery of Drug
Data on the Example of Adverse Reaction Prediction, BMC Bioinformatics, vol.
15 Suppl 6, p. S7, 2014. doi: 10.1186/1471-2105-15-S6-S7.
[28] Zhang F., Sun B., Diao X., Zhao W., Shu T.: Prediction of Adverse Drug Re-
actions Based on Knowledge Graph Embedding, BMC Medical Informatics and
Decision Making, vol. 21(1), p. 38, 2021. doi: 10.1186/s12911-021-01402-3.
[29] Zhao Y., Yi M., Tiwari R.C.: Extended Likelihood Ratio Test-Based Methods
for Signal Detection in a Drug Class with Application to FDA’s Adverse Event
Reporting System Database, Statistical Methods in Medical Research, vol. 27,
pp. 876–890, 2018.
[30] Zheng Y., Peng H., Zhang X., et al.: Predicting Adverse Drug Reactions of Com-
bined Medication from Heterogeneous Pharmacologic Databases, BMC Bioinfor-
matics, vol. 19(Suppl 19), p. 517, 2018. doi: 10.1186/s12859-018-2520-8.
[31] Zhou F., Khushi M., Brett J., Uddin S.: Graph Neural Network-Based Sub-
graph Analysis for Predicting Adverse Drug Events, Computers in Biology and
Medicine, vol. 183, p. 109282, 2024. doi: 10.1016/j.compbiomed.2024.109282.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Computer Science
This work is licensed under a Creative Commons Attribution 4.0 International License.
