Toward novel anti‐inflammatory <scp>PI3Kδ</scp> inhibitors: A virtual screening approach leveraging consensus pharmacophore models and multi‐algorithm docking

Abstract Phosphoinositide 3‐kinase delta (PI3Kδ) is a therapeutic target for autoimmune and inflammatory diseases. However, the development of selective inhibitors remains challenging due to the high structural homology among PI3K isoforms. This research sought to identify new PI3Kδ inhibitors using a comprehensive computational strategy that integrates pharmacophore modeling, molecular docking, and consensus scoring. First, common feature and receptor‐ligand pharmacophores were generated based on GSK2269557 and its crystal structure with PI3Kδ. The optimal common feature model (CF_10, AUC = 0.943) and the receptor–ligand model (RL_03, AUC = 0.942) demonstrated excellent discriminatory power. Then, four docking programs were benchmarked; AutoDock Vina and Glide XP exhibited superior pose reproduction and inhibitor classification (AUC &gt; 0.9). A consensus scoring function integrating pharmacophore fit values and docking scores was developed for virtual screening of the ChEMBL database, yielding several high‐ranking virtual hits, with CHEMBL2216838 identified as the top candidate. Finally, 300 ns molecular dynamics simulations revealed that CHEMBL2216838 forms stable hydrogen bonds with key residues Glu826/Val828 and induces a more rigid binding mode compared to GSK2269557, enhancing interactions with Phe912. This work establishes a robust computational framework for identifying novel PI3Kδ inhibitors and presents promising candidates for further experimental validation in the treatment of inflammation‐related diseases.