DECISION MAKING IN ANTI-CORONAVIRUS DRUG DISCOVERY: MATHEMATICAL MODELING AND VALUE OF INFORMATION ANALYSIS

20 2025 / 4 KPI Science News DOI: https://doi.org/10.20535/kpisn.2025.4.344350 UDC 519.6; 615.015.8 D.S. Horodetskyi1*, M.P. Smetiukh1,2, S.O. Soloviov1,2 National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Kyiv, Ukraine, 2 Shupyk National Healthcare University of Ukraine, Kyiv, Ukraine 1 Corresponding author: * DECISION MAKING IN ANTI-CORONAVIRUS DRUG DISCOVERY: MATHEMATICAL MODELLING AND VALUE OF INFORMATION ANALYSIS Background. The process of preclinical evaluation of antiviral medications typically involves multiple stages, each containing substantial uncertainties. Traditional methods for screening the compounds often lack structured means for optimising the decision-making and calculating the feasibility and risks of transitions between all of the stages. Thus, there appears to be a problem with the inefficient selection of promising antiviral molecules, which subsequently increases the probability of choosing suboptimal research trajectories. Objective. The paper aims to develop a computational framework for optimising of the transition between stages in preclinical antiviral testing. The system focuses on the integration of decision trees and Markov models in order to include effectiveness, risks and the value of additional information into assessment, supporting an in-depth planning of preclinical research pipelines. Methods. Experimental data from molecular docking, cytotoxicity CD50, and antiviral activity IC50 were used in a multi-stage evaluation system with CTI ≥ 4 being the criterion for progression into further stages. Decision trees provided the explicit rules for advancement of the compounds, while Markov models added context for building sequential strategies under uncertainty and quantified the feasibility of movement to the next stage. Value of information analysis added the assessment of the expected benefit of additional data. Results. The developed framework consistently produced reliable technical results. The decision used in CTI ≥ 4.0 prediction stage demonstrated a conservative classification pattern, correctly identifying compounds with high therapeutic potential while missing some effective candidates. The Markov model showed steadily increasing state values in docking, cytotoxicity, and antiviral testing phases that confirmed the growth of expected utility. Based on the findings acquired, the most effective solutions were identified for the ongoing investigation into antiviral assays, while the application of value of information analysis indicated that the largest gain occurred after antiviral activity testing, whereas the initial phases serve as filters. Conclusions. The study showed that both decision trees and Markov models capture different but complementary aspects of the preclinical evaluation process. Decision trees provide an interpretable set of rules that formalise how molecular docking and cytotoxicity measurement influence the progression of compounds, while their limited sensitivity at the CTI threshold highlighted the complexity of predicting the final success of the evaluated compounds. The Markov model simulations showed that the full three-stage pipeline is justified and that progression decisions are influenced by both uncertainty and experimental cost. The value of information analysis clarifies the importance of each stage, helping to emphasise the role of antiviral activity data. These findings support the integration of analytic methods for improving the structure, transparency and efficiency of antiviral preclinical research. Keywords: coronavirus; drug; preclinical evaluation; decision tree; Markov decision process; value of information. Introduction The optimisation of sequential decision-making in preclinical studies of antiviral compounds remains a highly relevant challenge due to the combination of uncertainty, high experimental costs, and limited predictability of candidate efficacy. At each stage of the preclinical pipeline – from in silico scree ning to cytotoxicity assessment and antiviral activity tests – researchers must make a series of decisions, Пропозиція для цитування цієї статті: Д.С. Городецький, М.П. Сметюх, С.О. Соловйов, «Прийняття рішень у процесі доклінічної розробки противірусних препаратів проти коронавірусу: математичне моделювання та аналіз цінності інформації», Наукові вісті КПІ, № 4, с. 20–30, 2025. doi: https://doi.org/10.20535/kpisn.2025.4.344350 Offer a citation for this article: D.S. Horodetskyi, M.P. Smetiukh, S.O. Soloviov, “Decision making in anti-coronavirus drug discovery: mathematical modelling and value of information analysis”, KPI Science News, No. 4, pp. 20–30, 2025. doi: https://doi.org/10.20535/kpisn.2025.4.344350 © The Autor(s). The article is distributed under the terms of the license CC BY 4.0 ПРИКЛАДНА МАТЕМАТИКА where an inaccurate early-stage choice leads to the loss of time, resources, and potentially promising compounds. This creates the need for systematic approaches capable of increasing the rationality and economic efficiency of the preclinical process. Despite significant progress in artificial intelligence, current research mainly improves individual steps of drug discovery rather than the full decision-making pipeline. Modern machine learning techniques demonstrate substantial advances in virtual screening, toxicity prediction, and target selection [1]. AI-based integration with organ-on-a-chip platforms and digital twins enhances the accuracy of pharmacokinetic and toxicological modelling [2]. Data-driven design of antiviral peptides using GANs, deep learning and explainable AI demonstrates strong potential for optimising candidate properties [3]. Studies of DHODH inhibitors highlight the complexity of translating promising in vitro results into clinical effects and emphasise the need for step-wise risk assessment [4]. Multi-omics deep learning pipelines accelerate early discovery and facilitate drug repositioning [5]. AI-based prediction of viral mutations supports personalised antiviral stra tegies and shows the sequential, dynamic nature of decision-making in virology [6]. AI-driven dereplication and classification of natural products further illustrate the need for structured transitions between preclinical stages [7]. However, these advances primarily address predictive accuracy rather than the principled optimisation of decisions across multiple stages. Current research lacks integrated mathematical frameworks that would: formalise transitions between preclinical stages, quantify risks and probabilities of success, incorporate the cost and value of information, and determine when experimental continuation is economi cally justified. Decision trees and Markov processes are rarely applied specifically to antiviral preclinical pipelines, leaving a methodological gap in modelling sequential choices under uncertainty. The study aims to develop and evaluate formalised approaches for optimising sequential deci (...truncated)