Enhancing drug discovery with AI: Predictive modeling of pharmacokinetics using Graph Neural Networks and ensemble learning

Abstract

Accurately predicting pharmacokinetic (PK) parameters such as absorption, distribution, metabolism, and excretion (ADME) is essential for optimizing drug efficacy, safety, and development timelines. Traditional experimental methods are often slow and expensive, driving the need for advanced AI-based approaches in PK modeling. This study compares cutting-edge machine learning models, including Graph Neural Networks (GNNs), Transformers, and Stacking Ensembles, against traditional models like Random Forest and XGBoost, using a dataset of over 10,000 bioactive compounds from the ChEMBL database. The Stacking Ensemble model achieved the highest accuracy (R² of 0.92, MAE of 0.062), outperforming GNNs (R² of 0.90) and Transformers (R² of 0.89). These AI models excelled in capturing complex molecular interactions and long-range dependencies, significantly improving PK predictions. The high accuracy achieved (R² = 0.92) by the Stacking Ensemble method indicates that AI models can streamline the drug discovery process by reducing costly in vivo experiments, enabling faster go/no-go decisions during preclinical evaluations, and ultimately accelerating the development of new thera-peutics. This reduction in time and cost could facilitate broader industry adoption of AI-driven PK modeling. Furthermore, Bayesian optimization was employed to fine-tune hyperparameters, further enhancing the performance and robustness of these predictive models.

Keywords

Pharmacokinetics (PK) / Machine learning models / Graph neural networks (GNNs) / Stacking ensemble / Drug discovery

References

[1] Uno Mizuki, Nakamaru Yuta, Yamashita Fumiyoshi. Application of machine learning techniques in population pharmacokinetics/pharmacodynamics modeling. Drug Metabol Pharmacokinet. 2024; 56: 101004. ISSN 1347-4367.

[2] Lai Yurong, Chu Xiaoyan, Di Li, et al. Recent advances in the translation of drug metabolism and pharmacokinetics science for drug discovery and development. Acta Pharm Sin B. 2022; 12 (6): 2751- 2777. ISSN 2211-3835

[3] Li Y, Meng Q, Yang M, et al. Current trends in drug metabolism and pharmacokinetics. Acta Pharm Sin B. 2019 Nov; 9 (6): 1113- 1144. Epub 2019 Oct 18. PMID: 31867160; PMCID: PMC6900561.

[4] Paul D, Sanap G, Shenoy S, Kalyane D, Kalia K, Tekade RK. Artificial intelligence in drug discovery and development. Drug Discov Today. 2021 Jan; 26 (1): 80- 93. Epub 2020 Oct 21. PMID: 33099022; PMCID: PMC7577280.

[5] Vora LK, Gholap AD, Jetha K, Thakur RRS, Solanki HK, Chavda VP. Artificial intelligence in pharmaceutical technology and drug delivery design. Pharmaceutics. 2023 Jul 10; 15 (7): 1916. PMID: 37514102; PMCID: PMC10385763.

[6] Patel Veer, Shah Manan. Artificial intelligence and machine learning in drug discovery and development. Intelligent Medicine. 2022; 2 (3): 134- 140. ISSN 2667-1026.

[7] Gangwal A, Ansari A, Ahmad I, et al. Generative artificial intelligence in drug discovery: basic framework, recent advances, challenges, and opportunities . Front Pharmacol. 2024 Feb 7; 15: 1331062. PMID: 38384298; PMCID: PMC10879372.

[8] Durap A. A comparative analysis of machine learning algorithms for predicting wave runup. Anthropocene Coasts. 2023; 6: 17.

[9] Sarker IH. Machine learning: algorithms, real-world applications and research directions. SN COMPUT. SCI. 2021; 2: 160.

[10] Nadkarni SB, Vijay GS, Kamath RC. Comparative study of random forest and gradient boosting algorithms to predict airfoil self-noise. Eng Proc. 2023; 59: 24.

[11] Xiong Jiacheng, Xiong Zhaoping, Chen Kaixian, Jiang Hualiang, Zheng Mingyue. Graph neural networks for automated de novo drug design. Drug Discov Today. 2021; 26 (6): 1382- 1393. ISSN 1359-6446.

[12] Jiménez-Luna J, Grisoni F, Schneider G. Drug discovery with explainable artificial intelligence. Nat Mach Intell. 2020; 2: 573- 584.

[13] Bullock Joseph, Luccioni Alexandra, Pham Katherine Hoffman, Lam Cynthia Sin Nga, Luengo-OrozVisan Miguel. Mapping the landscape of artificial intelligence applications against COVID-19. J Artif Intell Res. 2020; 69: 807- 845.

[14] Hua Liwen, Wang Danni, Wang Keran, et al. Design of tracers in fluorescence polarization assay for extensive application in small molecule drug discovery. J Med Chem. 2023; 66 (16): 10934- 10958. ISSN 1520-4804.

[15] Dostmohammadi Mahziyar, Zamani Pedram Mona, Hoseinzadeh Siamak, Astiaso Garcia Davide. A GA-stacking ensemble approach for forecasting energy consumption in a smart household: a comparative study of ensemble methods. J Environ Manag. 2024; 364: 121264. ISSN 0301-4797.

[16] Yang Siyun, Kar Supratik. Application of artificial intelligence and machine learning in early detection of adverse drug reactions (ADRs) and drug-induced toxicity. Artificial Intelligence Chemistry. 2023; 1 (2): 100011. ISSN 2949-7477.

[17] Nguyen V. Bayesian optimization for accelerating hyper-parameter tuning. In: 2019 IEEE Second International Conference on Artificial Intelligence and Knowledge Engineering (AIKE). . 2019: 302- 305. Sardinia, Italy.

[18] Wu Jia, Chen Xiu-Yun, Zhang Hao, Xiong Li-Dong, Lei Hang, Deng Si-Hao. Hyperparameter optimization for machine learning models based on bayesian optimizationb. Journal of Electronic Science and Technology. 2019; 17 (1): 26- 40. ISSN 1674-862X.

[19] Ogami C, Tsuji Y, Seki H, et al. An artificial neural network-pharmacokinetic model and its interpretation using Shapley additive explanations. CPT Pharmacometrics Syst Pharmacol. 2021 Jul; 10 (7): 760- 768. Epub 2021 May 27. PMID: 33955705; PMCID: PMC8302242.

[20] Gill J, Moullet M, Martinsson A, et al. Evaluating the performance of machinelearning regression models for pharmacokinetic drug-drug interactions. CPT Pharmacometrics Syst Pharmacol. 2023 Jan; 12 (1): 122- 134. Epub 2022 Nov 17. PMID: 36382697; PMCID: PMC9835131.

[21] Obrezanova Olga. Artificial intelligence for compound pharmacokinetics prediction. Curr Opin Struct Biol. 2023; 79: 102546. ISSN 0959-440X.

[22] Keutzer L, You H, Farnoud A, et al. On Behalf Of The Unite Tb Consortium. Machine learning and pharmacometrics for prediction of pharmacokinetic data: differences, similarities and challenges illustrated with rifampicin. Pharmaceutics. 2022 Jul 22; 14 (8): 1530. PMID: 35893785; PMCID: PMC9330804.

[23] Qureshi Rizwan, Irfan Muhammad, Gondal Taimoor Muzaffar, et al. AI in drug discovery and its clinical relevance. Heliyon. 2023; 9 (7): e17575. ISSN 2405-8440.

[24] Duan Feng-Lei, Duan Chun-Bao, Xu Hui-Lin, et al. AI-driven drug discovery from natural products. Advanced Agrochem. 2024; 3 (3): 185- 187. ISSN 2773-2371.

[25] Fan Jianing, Shi Shaohua, Xiang Hong, et al. J Chem Inf Model. 2024; 64 (8): 3080- 3092.

[26] Kiriiri GK, Njogu PM, Mwangi AN. Exploring different approaches to improve the success of drug discovery and development projects: a review. Futur J Pharm Sci. 2020; 6: 27.

[27] Mustapha Babajide, Ismail , Saeed Faisal. Bioactive molecule prediction using extreme gradient boosting. Molecules. 2016; 21 (8): 983.

[28] Chou WC, Lin Z. Machine learning and artificial intelligence in physiologically based pharmacokinetic modeling. Toxicol Sci. 2023 Jan 31; 191 (1): 1- 14. PMID: 36156156; PMCID: PMC9887681.

[29] Bargam B, Boudhar A, Kinnard C, et al. Evaluation of the support vector regression (SVR) and the random forest (RF) models accuracy for streamflow prediction under a data-scarce basin in Morocco. Discov Appl Sci. 2024; 6: 306.

[30] Cáceres EL, Tudor M, Cheng AC. Deep learning approaches in predicting ADMET properties. Future Med Chem. 2020; 12 (22): 1995- 1999.

[31] Ogami C, Tsuji Y, Seki H, et al. An artificial neural network-pharmacokinetic model and its interpretation using Shapley additive explanations. CPT Pharmacometrics Syst Pharmacol. 2021 Jul; 10 (7): 760- 768. Epub 2021 May 27. PMID: 33955705; PMCID: PMC8302242.

[32] Pugliese Raffaele, Regondi Stefano, Marini Riccardo. Machine learning-based approach: global trends, research directions, and regulatory standpoints. Data Science and Management. 2021; 4: 19- 29. ISSN 2666-7649.

[33] Cheng Yiming, Hu Hongxiang, Dong Xin, Xiaoran Hao, Li Yan. Exploring transformer model in longitudinal pharmacokinetic/pharmacodynamic analyses and comparing with alternative natural language processing models. J Pharmaceut Sci. 2024; 113 (5): 1368- 1375. ISSN 0022-3549.

[34] Shanmugasundar G, Vanitha M, Čep Robert, Kumar Vikas, Kalita Kanak, Ramachandran M. A comparative study of linear, random forest and AdaBoost regressions for modeling non-traditional machining. Processes. 2021; 9 (11): 2015.

[35] Salehin Imrus, Shamiul Islam Md, Saha Pritom, et al. AutoML: a systematic review on automated machine learning with neural architecture search. Journal of Information and Intelligence. 2024; 2 (1): 52- 81. ISSN 2949-7159.

[36] Chen W, Liu X, Zhang S, Chen S. Artificial intelligence for drug discovery: resources, methods, and applications. Mol Ther Nucleic Acids. 2023 Feb 18; 31: 691- 702. PMID: 36923950; PMCID: PMC10009646.

[37] Chen Haojie, Sloggy Matthew R, Dhiaulhaq Ahmad, et al. Boundary of ecosystem services: guiding future development and application of the ecosystem service concepts. J Environ Manag. 2023; 344: 118752. ISSN 0301-4797.

[38] Grogan S, Preuss CV. Pharmacokinetics. Treasure Island (FL): StatPearls Publishing; 2024 Jan [Updated 2023 Jul 30]. In: StatPearls [Internet] https://www.ncbi.nlm.nih.gov/books/NBK557744/..

[39] Yu RH, Cao YX. A method to determine pharmacokinetic parameters based on andante constant-rate intravenous infusion. Sci Rep. 2017 Oct 16; 7 (1): 13279. PMID: 29038495; PMCID: PMC5643313.

[40] Sayre RR, Wambaugh JF, Grulke CM. Database of pharmacokinetic time-series data and parameters for 144 environmental chemicals. Sci Data. 2020; 7: 122.

[41] Chang ED, Hogstrand C, Miller TH, Owen SF, Bury NR. The use of molecular descriptors to model pharmaceutical uptake by a fish primary gill cell culture epithelium. Environ Sci Technol. 2019 Feb 5; 53 (3): 1576- 1584. Epub 2019 Jan 15. PMID: 30589539; PMCID: PMC6503469.

[42] Gaulton A, Bellis LJ, Bento AP, et al. ChEMBL: a large-scale bioactivity database for drug discovery. Nucleic Acids Res. 2012 Jan; 40 (Database issue): D1100- D1107. Epub 2011 Sep 23. PMID: 21948594; PMCID: PMC3245175.

[43] Fu Yingpeng, Liao Hongjian, Lv Longlong. A comparative study of various methods for handling missing data in UNSODA. Agriculture. 2021; 11 (8): 727.

[44] Kamrul Hasan Md, Ashraful Alam Md, Roy Shidhartho, Dutta Aishwariya, Tasnim Jawad Md, Das Sunanda. Missing value imputation affects the performance of machine learning: a review and analysis of the literature (2010-2021). Inform Med Unlocked. 2021; 27: 100799. ISSN 2352-9148.

[45] Chikodili NB, Abdulmalik MD, Abisoye OA, Bashir SA. Outlier detection in multivariate time series data using a fusion of K-medoid, standardized euclidean distance and Z-score. In: Misra S, Muhammad-Bello B, eds. Information and Communication Technology and Applications. ICTA 2020. Cham: Springer; 2021:. Communications in Computer and Information Science; vol. 1350.

[46] Cannon M, Stevenson J, Kuzma K, et al. Normalization of drug and therapeutic concepts with Thera-Py. JAMIA Open. 2023 Nov 8; 6 (4): ooad093. PMID: 37954974; PMCID: PMC10637840.

[47] Khurana U, Samulowitz H, Turaga D. Feature engineering for predictive modeling using reinforcement learning. Proc AAAI Conf Artif Intell. 2018; 32 (1).

[48] Ma L, Zheng J. A polynomial based model for cell fate prediction in human diseases. BMC Syst Biol. 2017 Dec 21; 11 (Suppl 7): 126. PMID: 29322923; PMCID: PMC5770079.

[49] Huang HN, Chen HM, Lin WW, et al. Employing feature engineering strategies to improve the performance of machine learning algorithms on echocardiogram dataset. Digit Health. 2023 Oct 29; 58 (9): 20552076231207589. PMID: 37915794; PMCID: PMC10617266.

[50] Yin Yongjing, Lai Shaopeng, Song Linfeng, et al. An external knowledge enhanced graph-based neural network for sentence ordering. J Artif Intell Res. 2021; 70 (2021): 545- 566.

[51] Zhang Zehong, Chen Lifan, Zhong Feisheng, et al. Graph neural network approaches for drug-target interactions. Curr Opin Struct Biol. 2022; 73: 102327. ISSN 0959-440X.

[52] Yao R, Shen Z, Xu X, et al. Knowledge mapping of graph neural networks for drug discovery: a bibliometric and visualized analysis. Front Pharmacol. 2024 May 10; 15: 1393415. PMID: 38799167; PMCID: PMC11116974.

[53] Jiang Jian, Chen Long, Ke Lu, et al. A review of transformers in drug discovery and beyond. Journal of Pharmaceutical Analysis. 2024: 101081. ISSN 2095-1779.

[54] Aurpa TT, Sadik R, Ahmed MS. Abusive Bangla comments detection on Facebook using transformer-based deep learning models. Soc. Netw. Anal. Min. 2022; 12: 24.

[55] Lu M, Hou Q, Qin S, et al. A stacking ensemble model of various machine learning models for daily runoff forecasting. Water. 2023; 15: 1265.

[56] Abbas S, Sampedro GA, Abisado M, Almadhor AS, Kim T-H, Zaidi MM. A novel drug-drug indicator dataset and ensemble stacking model for detection and classification of drug-drug interaction indicators. IEEE Access. 2023; 11: 101525- 101536.

[57] Shanmugasundar G, Vanitha M, Čep Robert, Kumar Vikas, Kalita Kanak, Ramachandran M. A comparative study of linear, random forest and AdaBoost regressions for modeling non-traditional machining. Processes. 2021; 9 (11): 2015.

[58] Mustapha Babajide, Ismail , Saeed Faisal. Bioactive molecule prediction using extreme gradient boosting. Molecules. 2016; 21 (8): 983.

[59] Fu Yingpeng, Liao Hongjian, Lv Longlong. A comparative study of various methods for handling missing data in UNSODA. Agriculture. 2021; 11 (8): 727.

[60] Khurana U, Samulowitz H, Turaga D. Feature engineering for predictive modeling using reinforcement learning. Proc AAAI Conf Artif Intell. 2018; 32 (1).

[61] Belakaria Syrine, Deshwal Aryan, Rao Janardhan. Output space entropy search framework for multi-objective bayesian optimization. J Artif Intell Res. 2021; 72: 667- 715.

[62] Jierula A, Wang S, Oh T-M, Wang P. Study on accuracy metrics for evaluating the predictions of damage locations in deep piles using artificial neural networks with acoustic emission data. Appl Sci. 2021; 11: 2314.

RIGHTS & PERMISSIONS

The Authors. Publishing services by Elsevier B.V. on behalf of Higher Education Press and KeAi Communications Co. Ltd.