combined

CombinedFilterQuery dataclass

Query object to apply combined filtering.

Parameters:

Name	Type	Description	Default
min_confidence	Annotated[float, Parameter(validator=Number(lte=100, gte=0))]	Minimal confidence (plDDT) for AlphaFold structures to pass the filter.	70.0
min_residues	NonNegativeInt	Min residues in chain A.	0
max_residues	PositiveInt	Max residues in chain A.	10000000
min_geometry_quality	float	Minimal geometry quality score to pass the filter.	50.0
min_sequence_identity	NormFloat	Minimum sequence identity ratio to the Uniprot sequence for a structure to be passed. If not set then discards structures that are not fully identical to the Uniprot sequence. For example if set to 0.8 then structures that have sequence identity below 0.8 are discarded.	1.0
top_uniprot_cluster	NonNegativeInt	Maximum number of files to keep for structures per cluster per Uniprot accession. Alphafold structures are excluded from this limit.	1000
top_non_uniprot	NonNegativeInt	Maximum number of files to keep for structures without Uniprot accession.	0

CombinedFilterResult dataclass

Result of combined filtering.

Compatible with SortableStructure protocol.

Parameters:

Name	Type	Description	Default
input_file	Path | None	Path to the input structure file.	None
pdb_id	str | None	PDB ID of the structure.	None
metadata	StructureMetadata | None	Structure metadata.	None
high_confidence_residues_count	int | None	Number of residues with high confidence (plDDT) for AlphaFold structures.	None
geometry_quality	float | None	Geometry quality score.	None
passed	bool	Whether the structure passed the filter.	False
reason	str | None	Reason for failure if the structure did not pass the filter.	None
output_file	Path | None	Path to the output structure file if the structure passed the filter.	None

CombinedPartitions dataclass

Partitions of structures and results.

Used internally by the combined_filter function.

Attributes:

Name	Type	Description
uniprot_with_resolution	list[ResolutionOrGeometryQualityClusterableStructure]	Structures with UniProt accession and resolution.
uniprot_with_geometry_quality	list[ResolutionOrGeometryQualityClusterableStructure]	Structures with UniProt accession and geometry quality.
resolution_only	list[CombinedFilterResult]	Structures with resolution but no UniProt accession.
geometry_quality_only	list[CombinedFilterResult]	Structures with geometry quality but no UniProt accession.
processed	list[CombinedFilterResult]	Results of structures that have been processed.

ResolutionOrGeometryQualityClusterableStructure dataclass

Structure that can be clustered.

Compatible with ClusterableStructure protocol.

Used internally by the combined_filter function.

Parameters:

Name	Type	Description	Default
input_file	Path	Path to the input structure file.	required
metadata	StructureMetadata	Structure metadata.	required
geometry_quality	float | None	Geometry quality score.	None

combined_filter(input_files, scores, filters, output_dir, copy_method='hardlink', scheduler_address=None)

Combined filter for PDB/mmCIF files.

Rules:

• All non-AlphaFold structures are filtered by number of residues in chain A. See protein-quest filter residue --help for details.
• All non-AlphaFold structures with Uniprot accession are filtered by sequence identity.
• AlphaFold structures are filtered by confidence (plDDT) and afterwards by number of residues in chain A. See protein-quest filter confidence --help for details.
• Structures with uniprot accession and resolution are filtered by grouping/clustering and sorting cluster members by resolution. See protein-quest filter resolution --help for details.
• Structures with Uniprot accesion and without resolution are filtered by grouping/clustering and sorting cluster members by PDBe quality scores.
• Structures without Uniprot accession and with resolution are filtered/sorted by resolution.
• Structures without Uniprot accession and without resolution are filtered by PDBe quality scores.
• Structures without Uniprot accession, without resolution and without PDBe quality scores are discarded.

flowchart TD
    A[Input PDB/mmCIF files] --> B{AlphaFold structure?}

    B -->|Yes| C[Filter by confidence plDDT]
    C --> D[Filter by residues in chain A]

    B -->|No| F[Filter by residues in chain A]
    F --> U[Filter by sequence identity]
    U --> G{UniProt accession?}

    G -->|Yes| H{Resolution available?}
    H -->|Yes| I[Group by UniProt accession and cluster by residue ranges]
    I --> J[Sort cluster members by resolution]
    J --> K[Keep up to top_uniprot_cluster per cluster]

    H -->|No| L[Group by UniProt accession and cluster by residue ranges]
    L --> M[Sort cluster members by PDBe quality]
    M --> N[Keep up to top_uniprot_cluster per cluster]

    G -->|No| O{Resolution available?}
    O -->|No| P[Sort by PDBe quality]
    P --> R[Select up to top_non_uniprot entries]
    O -->|Yes| S[Sort by resolution]
    S --> T[Select up to top_non_uniprot entries]

    D --> Q[Write output structure files]
    K --> Q
    N --> Q
    R --> Q
    T --> Q

Parameters:

Name	Type	Description	Default
input_files	list[Path]	List of input PDB/mmCIF files.	required
scores	dict[str, Scores]	Dictionary of PDB IDs to PDBe scores.	required
filters	CombinedFilterQuery	Combined filter query.	required
output_dir	Path	Directory to write output structure files.	required
copy_method	CopyMethod	Method to copy files to output directory.	'hardlink'
scheduler_address	SchedulerAddress	Address of the Dask scheduler for parallel processing.	None

Returns:

Type	Description
list[CombinedFilterResult]	List of results.