MHC-Diff: Evaluation Data for Peptide–MHC Structure Prediction

Evaluation data accompanying MHC-Diff: Fast and Accurate Peptide–MHC Structure Prediction via an Equivariant Diffusion Model. This dataset contains the per-case evaluation data behind the paper's benchmark tables and figures: for every X-ray test case, the crystal target and MHC-Diff's ten predicted Cα conformations, together with per-case Cα-RMSD values for MHC-Diff and the baseline methods. Data covers the 8K (HLA-A*02:01 9-mer) and 100K (multi-allele generalization) benchmarks, plus the G-domain clustering used to define held-out test sets. See README.md for the full mapping of files to each table and figure. Code: https://github.com/X-lab-3D/MHC-Diff