Explainable Artificial Intelligence-Guided Virtual Screening of Supramolecular Organic Assemblies for Multi-Target Alzheimer’s Disease Therapy

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder characterized by β-amyloid plaque accumulation, tau protein aggregation, and neuronal loss. Current therapies provide limited symptomatic relief, emphasizing the need for effective multi-target therapeutic strategies. Supramolecular organic assemblies, including cyclodextrins, calixarenes, and cucurbiturils, exhibit unique host–guest interactions and structural versatility, making them promising candidates for drug discovery. This study proposes an integrated in silico framework combining explainable artificial intelligence (XAI), cheminformatics, molecular docking, and ADMET prediction to identify potential supramolecular candidates against AD. A curated dataset of Alzheimer’s-related compounds and supramolecular scaffolds will be analyzed using molecular descriptors and molecular fingerprints to develop machine learning models targeting β-secretase (BACE1), acetylcholinesterase (AChE), and glycogen synthase kinase-3β (GSK-3β). Explainable AI techniques, including SHAP and LIME, will be employed to interpret model predictions and identify key molecular features associated with biological activity. Top-ranked candidates will undergo molecular docking to evaluate target-binding interactions, followed by computational ADMET analysis to assess pharmacokinetic and toxicity profiles. The proposed workflow integrates predictive modelling, explainability, and molecular-level validation, providing a transparent and reproducible strategy for accelerating the computational discovery and rational design of multi-target supramolecular anti-Alzheimer agents. Identified candidates may serve as promising leads for subsequent experimental validation and therapeutic development. By integrating complementary computational approaches, the study aims to prioritize compounds with favourable target affinity, physicochemical properties, and safety characteristics, while providing mechanistic insights into structure–activity relationships and facilitating the design of novel supramolecular systems with improved therapeutic potential.