plinder.data.linked_apo

class plinder.data.linked_apo.LinkedApoSelectionConfig(max_per_system: int = 5, min_pocket_fident: int = 95, min_protein_fident_weighted_sum: int = 95, min_protein_fident_qcov_weighted_sum: int = 80, min_protein_lddt_weighted_sum: int = 20)

Bases: object

Similarity thresholds and the maximum apo links per holo system.

max_per_system: int = 5

min_pocket_fident: int = 95

min_protein_fident_weighted_sum: int = 95

min_protein_fident_qcov_weighted_sum: int = 80

min_protein_lddt_weighted_sum: int = 20

plinder.data.linked_apo.ligand_holo_chain_keys(annotation: DataFrame | str | Path) → DataFrame

Return receptor-chain keys bound to proper ligands.

plinder.data.linked_apo.build_apo_candidate_manifest(entry_chains: DataFrame | str | Path, *, biounit_chains: DataFrame | str | Path, entry_metadata: DataFrame | str | Path, annotation: DataFrame | str | Path) → DataFrame

Build one scored target row per apo protein chain.

A protein chain is apo when it is not ligand-like, is not bound to a proper ligand, and does not share an entity with a proper-ligand receptor chain in its entry. Protein-interface membership is independent of this ligand-relative definition. If a chain occurs in several biological assemblies, the assembly with the cleanest chain-local ligand environment is retained.

plinder.data.linked_apo.select_linked_apo_structures(protein_scores: DataFrame | str | Path, *, annotation: DataFrame | str | Path, candidates: DataFrame | str | Path, config: LinkedApoSelectionConfig | None = None) → DataFrame

Associate PLINDER holo systems with matching deposited apo chains.

A candidate must satisfy every required metric for every proper ligand pocket in the holo system. Candidates from the holo entry itself are excluded. Passing candidates prefer no chain-local ligand contacts, then ion-only contacts, artifact contacts, and other ligand contacts before experimental resolution and similarity.

plinder.data.linked_apo.write_linked_apo_structure_table(protein_scores: DataFrame | str | Path, *, annotation: DataFrame | str | Path, candidates: DataFrame | str | Path, output_path: str | Path, config: LinkedApoSelectionConfig | None = None, scratch_dir: str | Path | None = None, threads: int = 1, memory_limit: str = '7GB') → Path

Select linked apo structures and write the compact release table.