Comparative Analysis of AlphaFold2, AlphaFold3 and ESMFold in Chimeric Antigen Receptor Structure Prediction for Pancreatic Cancer Immunotherapy

Immunotherapy is a pioneer therapeutic approach with auspicious results in various cancer types. Pancreatic cancer is one of the deadliest types with no cellular immunotherapies currently approved and available for it. Our research aims to investigate the in-silico modeling and evaluation of a novel Chimeric Antigen Receptor (CAR) protein for Pancreatic Cancer Immunotherapy. Protein structure prediction algorithms can help toward this direction by precisely simulating such proteins, and further analyze and evaluate them with other computational methods, i.e., Molecular Docking and Molecular Dynamics (MD) simulations. In this study, two powerful protein structure prediction algorithms named AlphaFold2 (Nobel Prize in Chemistry 2024) and ESMFold are compared in the context of CAR prediction and its in-silico evaluation with Molecular Docking and MD software. The comparison is further enhanced through the integration of the state-of-the-art AlphaFold3 algorithm. The significant outcome of this computational research study is that ESMFold performs better than its competitors, thus revealing that Protein Language Models (PLMs) have notable and promising potential in the field of Computational Protein Design, especially for novel immunotherapeutic proteins.