Exploration of potential drugs in treatment of hantavirus pulmonary syndrome using graph-theoretical QSPR modeling

Hantavirus Pulmonary Syndrome (HPS) is a severe rodent-borne viral disease that can rapidly progress from influenza-like symptoms to life-threatening cardiopulmonary failure, while therapeutic options remain limited. This study develops a quantitative structure–property relationship (QSPR) framework for a curated set of fourteen HPS-related drug candidates using key physicochemical endpoints. Molecular structures are modeled as graphs, and degree-based topological descriptors are systematically derived via the M-polynomial techniques. Linear, quadratic, power, and logistic regression models are fitted to quantify structure–property relations, and model quality is assessed using R, $$R^{2}$$ , RMSE, SE, and F-statistics, with internal predictivity evaluated through leave-one-out cross-validation (LOOCV) using $$Q^{2}$$ . Results indicate that selected degree-based descriptors capture meaningful variation in the considered physicochemical properties and support reproducible, computation-driven screening of candidate compounds. In future directions, integrating pharmacokinetic and toxicity-oriented computational assessment is expected to further enhance the practical relevance of the proposed framework.