17 lines (14 loc) · 652 Bytes

Protein Ideal — BM5.5 Relaxation Benchmark

What This Is

Dataset preparation and benchmark metadata for BM5.5 (257 protein-protein complexes). Sister repo to Protein_Relax_Pipeline.

Key Facts

257 complexes, 605 chains, 122,966 total residues
119 homo-multimeric targets deduplicated to unique sequences
Crystal-derived FASTAs (not UniProt canonical) for accurate RMSD
41 targets with expression tags requiring sequence cleaning
4 non-standard entries: BAAD, BOYV, BP57, CP57

Key Files

PROJECT_STATUS.md — completion tracking
COMPARISON.md — benchmark comparison data
NOTEBOOK.md — development journal and decisions