Protein kinases play a central role in regulating cellular signalling pathways, and their dysregulation is strongly associated with cancer, inflammatory disorders, and other diseases. Consequently, protein kinase inhibitors have emerged as a major class of targeted therapeutics. However, traditional drug discovery approaches remain time-consuming, costly, and limited in exploring the vast chemical and biological space. In recent years, artificial intelligence (AI) and machine learning (ML) techniques have revolutionized the development of protein kinase inhibitors by enabling faster, more accurate, and cost-effective discovery pipelines. This review highlights the integration of AI-based approaches, including quantitative structure–activity relationship (QSAR) modelling, molecular docking, molecular dynamics simulations, and deep learning frameworks, in kinase inhibitor design and optimization. Advanced methods such as 3D convolutional neural networks, network-based drug repurposing, and hybrid structure-based and data-driven models have significantly improved target prediction, binding affinity estimation, and off-target profiling. Additionally, AI-driven platforms facilitate the identification of novel kinase targets and enable drug repositioning strategies, thereby accelerating therapeutic development. Despite these advancements, challenges such as data quality, model interpretability, and generalizability remain critical barriers to widespread implementation. Future perspectives emphasize the integration of multi-omics data, explainable AI, and collaborative databases to enhance predictive performance and translational success. Overall, AI-based approaches are transforming kinase inhibitor discovery, offering promising opportunities for precision medicine and next-generation drug development.
Anu et al. (Sat,) studied this question.