plankton package
Submodules
plankton.graph module
- class plankton.graph.SpatialGraph(sdata, n_neighbors=10)
Bases:
objectSpatialGraph: Container object for all KNN-graph based operations on spatial data sets.
- Parameters
sdata (plankton.SpatialData) – The spatial data source
n_neighbors (int, optional) – number of nearset neighbors to infer, defaults to 10
- property distances
distances: Returns an array of inter-spot distances, which each entry encoding the distance to a defined neighbor.
- Returns
2d array of distances to the data points of self.neighbors.
- Return type
numpy.Array
- knn_entropy(n_neighbors=4)
- map_and_umap(color_prop=None, scalebar=True, cmap='jet', **kwargs)
map_and_umap: Plots a side-by-side representation of the available UMAP- and coordinate data, with styling arguments passed to both plotting functions.
- Parameters
color_prop (_type_, optional) – Property to color the individual markers by. Needs to be a column in self.sdata.var, defaults to None
scalebar (bool, optional) – Whether to display a scalebar in the coordinate representation, defaults to True
- property neighbor_types
neighbor_types: Returns an array of gene_ids for the spots listed in self.neighbors
- Returns
2d array of gene_ids.
- Return type
numpy.Array
- property neighbors
neighbors: Returns an array of neighbors, where each entry encoding the row position of a point in the data set.
- Returns
2d array of n-neighbors per sample data point
- Return type
numpy.Array
- plot_entropy(n_neighbors=4)
- plot_umap(color_prop='genes', text_prop=None, text_color_prop=None, c=None, color=None, color_dict=None, text_distance=1, thlds_text=(1.0, 0.0, 0.0), text_kwargs={}, legend=False, **kwargs)
- run_umap(bandwidth=1, kernel=None, metric='euclidean', zero_weight=1.0, cutoff=None, random_state=None, *args, **kwargs)
run_umap: Creates a UMAP representation of recurring local contexts in the source data.
- Parameters
bandwidth (int, optional) – Bandwidth of the default Gaussian kernel used to build local environment models, defaults to 1
kernel (callable, optional) – Callable function to generate a model of the local spatial context for each spot’s KNN graph. If none is provided, the default Gaussian is used. Defaults to None
metric (str, callable, optional) – Kernel description used as a UMAP parameter, defaults to ‘euclidean’
zero_weight (float, optional) – Regularization parameter that adds information of each spot’s gene label to its local context model. High value encourages spots to form clusters with spots of their own gene label, defaults to 1
random_state (int) – Random state to initialize UMAP and FastICA.
- property umap
umap: Returns the coordinates of the UMAP representation of the source data.
- Returns
UMAP coordinates
- Return type
numpy.Array
- property umap_0
- property umap_1
- umap_js(color_prop='c_genes')
umap_js: A javascript-based display and selection function for the spatial source data and generated UMAP embedding. Can be used to understand UMAP agglomerations by displaying their gene compositions and projecting them back onto the original source data coordinates.
- Parameters
color_prop (str, optional) – Property by which to color the spots in the scatter plots of the source data coordinates as well as the UMAP representation. Needs to be a column in sdata.var, defaults to ‘c_genes’
- Returns
plotly HTML/javascript widget that displays in the jupyter notebook.
- Return type
plotly.Widget
- update_knn(n_neighbors, re_run=True)
update_knn: Updates the KNN representation that gives rise to self.neighbors, self.distances and self.neighbor_types. This does not happen automatically after slicing, so it should be performed regularity before any graph-based analysis. Can be memory-intensive.
- Parameters
n_neighbors (int) – Number of neighbors for each neighbor graph.
re_run (bool, optional) – re-run, or use current representation, defaults to True
plankton.pixelmaps module
plankton.plankton module
plankton.stats module
- plankton.stats.co_cooccurrence(sdata_0, sdata_1, resolution=5.0, max_radius=400, linear_steps=5, category=None)
co_occurrence _summary_
- Parameters
sdata (SpatialData) – SpatialData - all spots contained in sdata will be analysed.
resolution (float, optional) – Smallest resolution of the co-occurence model in um, defaults to 5.
max_radius (float, optional) – Largest radius of the co-occurrence model in um, defaults to 400.
linear_steps (int, optional) – Number of linear steps to model local heterogeniety. Afterwards, distance bins get wider to save computational resources, defaults to 5
category (str, optional) – Category to model the co-occurrence of. Must be a column in ‘sdata’ with dtype ‘category’. When given ‘None’, the algorithm defaults to gene labels.
- Returns
_description_
- Return type
_type_
- plankton.stats.co_occurrence(sdata, resolution=5.0, max_radius=400, linear_steps=5, category=None)
co_occurrence _summary_
- Parameters
sdata (SpatialData) – SpatialData - all spots contained in sdata will be analysed.
resolution (float, optional) – Smallest resolution of the co-occurence model in um, defaults to 5.
max_radius (float, optional) – Largest radius of the co-occurrence model in um, defaults to 400.
linear_steps (int, optional) – Number of linear steps to model local heterogeniety. Afterwards, distance bins get wider to save computational resources, defaults to 5
category (str, optional) – Category to model the co-occurrence of. Must be a column in ‘sdata’ with dtype ‘category’. When given ‘None’, the algorithm defaults to gene labels.
- Returns
_description_
- Return type
_type_
- plankton.stats.get_histograms(sdata, category=None, resolution=5, mins=None, maxs=None)
- plankton.stats.mor_normalize(stats1, stats2)
mor_normalize _summary_
- Parameters
stats1 (_type_) – _description_
stats2 (_type_) – _description_
- plankton.stats.ripleys_k(sdata, resolution=5, max_radius=400, linear_steps=5, category=None)