clustering
Cluster 3D-beacons UniProt summaries by PDBe residue-range overlap.
PDBe Overviews are clustered and pruned with the source-agnostic filter_structures_on_clustered_resolution. Overviews from other providers (AlphaFold, SWISS-MODEL, ...) pass through unchanged.
Attributes:
| Name | Type | Description |
|---|---|---|
PDBE_PROVIDER_RESPONSE |
Provider name used in the Overview summaries to identify entries provided by PDBe. |
OverviewClusterableEntry
dataclass
Adapter exposing the ClusterableStructure protocol.
Wraps an Overview so the generic clustering algorithm can consume it.
cluster_overviews_per_uniprot(summaries, top)
Cluster PDBe Overviews of each UniProt summary and keep at most top per accession cluster.
Overviews from providers other than PDBe are passed through unchanged and appended after the pruned PDBe Overviews.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
summaries
|
list[UniprotSummary]
|
UniProt summaries to prune. |
required |
top
|
int
|
Maximum number of PDBe Overviews to retain per UniProt accession cluster. |
required |
Returns:
| Type | Description |
|---|---|
list[UniprotSummary]
|
New list of UniprotSummary with PDBe Overviews pruned. |
Raises:
| Type | Description |
|---|---|
ValueError
|
If |