We curated a dataset of 186,685 MCP–protein complexes (mcpp_dataset.tar.gz) starting from 641 protein–MCP complexes from the RCSB PDB using a “mutate-then-relax” strategy:
Mutation:MCPs were randomly mutated at 1 to 8 sites using 213 distinct amino acids.
Relaxation:Mutated complexes were relaxed using FastRelax in Rosetta, which iteratively performs side-chain packing and… See the full description on the dataset page:
https://huggingface.co/datasets/Willete3/mcpp-dataset.
