grassp.tl.calculate_interfacialness_score

calculate_interfacialness_score(data, compartment_annotation_column, neighbors_key=None, obsp=None, exclude_category=None)

Calculate interfacialness scores for proteins based on their neighborhood annotations.

For each protein, examines its nearest neighbors and calculates a modified Jaccard coefficient between the two most frequent compartment annotations in the neighborhood. This provides a measure of how “interfacial” a protein is between different cellular compartments.

Parameters:

data AnnData

Annotated data matrix with proteins as observations (rows)

compartment_annotation_column str

Column in data.obs containing compartment annotations

neighbors_key Optional[str] (default: None)

Key for neighbors in data.uns. If not specified, will look for neighbors in obsp

obsp Optional[str] (default: None)

Key for neighbors in data.obsp. Only used if neighbors_key not specified

exclude_category Union[str, List[str], None] (default: None)

Category or list of categories to exclude from the analysis

Return type:

AnnData

Returns:

data Original AnnData object with added columns in .obs: - jaccard_score: Modified Jaccard coefficient measuring interfacialness - jaccard_d1: Number of neighbors with most frequent annotation - jaccard_d2: Number of neighbors with second most frequent annotation - jaccard_k1: Most frequent compartment annotation - jaccard_k2: Second most frequent compartment annotation