Fast and accurate modeling of TCR-peptide-MHC complexes using tFold-TCR 1. tFold-TCR introduces a high-throughput, end-to-end method for predicting TCR-peptide-MHC (TCR-pMHC) complex structures, overcoming challenges in TCR-pMHC interactions and enabling rapid and precise 3D structure modeling. 2. The model leverages a specially trained protein language model (PLM) that extracts intra- and inter-chain residue contact information, eliminating the need for laborious multiple sequence alignment (MSA) searches and achieving remarkable speed improvements. 3. Compared to existing methods like AlphaFold-3, tFold-TCR shows a 30.7% increase in prediction accuracy, achieving fast and reliable results in seconds, significantly enhancing computational efficiency. 4. The tFold-TCR model was used to construct TCRStructDB, the largest database of TCR-pMHC structures, including 2.2 million TCRs, 0.8 million pMHCs, and 45,000 TCR-pMHC complexes, providing unprecedented insights into immune system interactions. 5. With the ability to predict unliganded TCRs, unbound pMHCs, and complex TCR-pMHC structures, tFold-TCR advances our understanding of the immune system, supporting cancer immunotherapy and improving vaccine development. 6. Despite superior performance on TCRs and MHC-I interactions, the model's prediction of MHC-II complexes lags behind AlphaFold-3, indicating challenges due to the structural complexities of MHC-II molecules. 7. The platform allows large-scale structural predictions, offering a robust tool for drug discovery, vaccine design, and personalized immunotherapies, and is available via the online TCRStructDB database. 8. tFold-TCR sets a new standard in modeling TCR-pMHC complexes, providing a critical resource for advancing immunological research and therapeutic strategies. @Jamie_Rossjohn @RossjohnLab @supercs08 @aaguaijie 💻Code: github.com/TencentAI4S/tfold 📜Paper: biorxiv.org/content/10.1101/… #Immunotherapy #TCRModeling #CancerResearch #Bioinformatics #MachineLearning