plinder.core.scores.protein

Query compact chain-level protein similarities.

plinder.core.scores.protein.query_chain_overlap(query_entry: str, query_chain: str, target_entry: str, target_chain: str, *, kind: Literal['pocket', 'interface'], source: Literal['foldseek', 'mmseqs'] | None = None, search_db: str = 'holo', release: PlinderRelease | None = None) → DataFrame

Count aligned pocket or interface residues between two protein chains.

Chain IDs are label asym IDs. Residues are collected across all ligand pockets or all protein interfaces involving each chain. One row is returned per available search backend.

plinder.core.scores.protein.query_protein_similarity(*, columns: list[str] | None = None, filters: list[tuple[str, str, Any]] | list[list[tuple[str, str, Any]]] | None = None, release: PlinderRelease | None = None, include_monomers: bool = False) → DataFrame

Query chain-level Foldseek and MMseqs similarities in integer percent.

Set include_monomers to include queries from protein chains outside ligand pockets and protein interfaces.