The image container¶
The image container wraps methods that needs to run on multiple images. For example, clustering neighborhoods is typically performed on all samples simultaneously.
- class spatialproteomics.image_container.ImageContainer.ImageContainer(sprot_dict: Dict[str, Dataset])¶
This class is used to store multiple SpatialProteomics objects and perform operations on them.
- compute_neighborhoods(neighborhood_method: str = 'radius', radius=100, knn=10, k=5, overwrite: bool = False, seed: int = 0)¶
Compute neighborhoods for spatial proteomics objects using the specified method and perform clustering.
- Parameters:
neighborhood_method (str, optional) – The method to use for computing neighborhoods. Must be one of ‘radius’, ‘knn’, or ‘delaunay’. Default is ‘radius’.
radius (int, optional) – The radius to use for the ‘radius’ neighborhood method. Default is 100.
knn (int, optional) – The number of nearest neighbors to use for the ‘knn’ neighborhood method. Default is 10.
k (int, optional) – The number of clusters to form using K-Means clustering. Default is 5.
overwrite (bool, optional) – Whether to overwrite existing neighborhoods in the objects. Default is False.
seed (int, optional) – The random seed to use for K-Means clustering. Default is 0.
- Returns:
A dictionary of spatial proteomics objects with computed neighborhoods and clusters.
- Return type:
dict
- Raises:
ValueError – If neighborhoods are already present in the objects and overwrite is set to False. If neighborhood_method is not one of ‘radius’, ‘knn’, or ‘delaunay’. If there is an error in the clustering process.
- get_neighborhood_composition(standardize: bool = True) DataFrame¶
Get the composition of neighborhoods across all objects in the ImageContainer.
- Parameters:
standardize (bool, optional) – Whether to standardize the composition of neighborhoods. Default is True.
- Returns:
A DataFrame containing the composition of neighborhoods across all objects in the ImageContainer.
- Return type:
pd.DataFrame