An integrated machine learning and chemical space network approach for the design of potent epigenetic HDAC6 inhibitors for targeting neurological disorders.

Histone deacetylase 6 (HDAC6) is increasingly recognized as a key regulator of cytoskeletal dynamics and intracellular transport in both neurodevelopmental and neurodegenerative disorders. Despite the large number of HDAC6 inhibitors reported, only a few have shown effective in vivo activity in neurological disease models. In this context, machine-learning (ML) approaches offer a powerful strategy for extracting chemical, physical, and biological features from a large and complex dataset of 4307 HDAC6 inhibitors. The current research aimed to develop high-quality ML-based classification frameworks to identify pivotal structural fingerprints of HDAC6 inhibitors. Additionally, chemical space network (CSN) construction, scaffold diversity exploration, and matched molecular series analysis were performed to provide an extensive computational investigation of HDAC6 inhibitors. The research highlighted the significance of various scaffolds, which may play a promising role in HDAC6 inhibition. This integrated computational strategy offers a potent platform for interpreting and predicting HDAC6 inhibitory activity, thereby facilitating lead optimization and the rational design of novel therapeutics.