QSAR-guided pharmacophore modeling and subsequent virtual screening identify novel TYK2 inhibitor

Tyrosine Kinase 2 (TYK2) inhibition is of potential therapeutic value for treating autoimmune diseases. An elaborate ligand-based computational workflow was employed to explore structural requirements for TYK2 inhibition. Genetic function algorithm (GFA) was coupled to k-nearest neighbor (kNN) and multiple linear regression (MLR) analyses to search for predictive QSAR models based on optimal pharmacophore(s)/physicochemical descriptors combinations. QSAR-selected pharmacophores were validated by receiver operating characteristic (ROC) curve analysis and by comparison with crystallographic structures of known inhibitors complexed within the TYK2 binding pocket. Optimal QSAR models and their associated pharmacophore hypotheses were used to identify new TYK2 inhibitory leads retrieved from the National Cancer Institute (NCI) structural database. The most potent hit exhibited experimental anti-TYK2 IC50 of 7.1 µM.