Shukla_2026_Mol.Divers__

Alzheimer's disease (AD) is a chronic neurodegenerative disorder and a leading cause of dementia worldwide, characterized by progressive cognitive and memory decline. The impaired cholinergic neurotransmission in the brain is a major pathological feature of AD. Therefore, enhancing cholinergic function is a key therapeutic strategy for its management. In this study, a dataset of 111 phthalimide derivatives with experimental anti-AD activity was collected from various literature. Further, quantitative structure-activity relationship (QSAR) modelling was performed to determine key structural features governing acetylcholinesterase (AChE) inhibitory activity of phthalimide derivatives. Furthermore, Arithmetic Residuals in K-Groups Analysis (ARKA) was employed to develop ARKA and hybrid ARKA-RASAR models, that enabled the interpretation of QSAR descriptors in different response ranges and identification of similarity between closely related compounds as well as improving the predictive reliability of the QSAR model. All the developed models showed strong internal predictivity (R(2) = 0.75-0.79 and Q(2)(LOO) = 0.68-0.76) and good external predictivity (Q(2)(F1) and Q(2)(F2) = 0.54-0.59). The best model (hybrid ARKA-RASAR) was applied to virtually screen 79,947 phthalimide derivatives downloaded from the PubChem database. Compounds with IC50 values below 240 nM were shortlisted, resulting in 27 candidates that were further evaluated through molecular docking and binding free energy analysis. The top five hits, selected based on favourable docking scores and interactions, were subsequently assessed for ADMET studies. Among them, compound Ph1 was found to have favourable ADME and toxicity profile. This integrated computational study highlighted compound Ph1 as a promising AChE inhibitor for the management of AD.