spapros.se.ProbesetSelector.plot_clf_genes

ProbesetSelector.plot_clf_genes(basis='X_umap', celltypes=None, till_rank=1, importance_th=None, add_marker_genes=True, neighbors_params={}, umap_params={}, **kwargs)

Plot umaps of selected genes needed for cell type classification of each cell type.

Parameters:

  • basis (str) – Name of the obsm embedding to use.

  • celltypes (Optional[List[str]]) – Subset of cell types for which to plot decision genes. If None, celltypes is used.

  • till_rank (Optional[int]) – Plot decision genes only up to the given tree rank of the probeset list.

  • importance_th (Optional[float]) – Only plot genes with a tree feature importance above the given threshold.

  • add_marker_genes (bool) – Whether to add subplots for marker genes from marker_list for each celltype. TODO: what about cell types that only occur in the marker list?

  • neighbors_params (dict) – Parameters for sc.pp.neighbors(). Only applicable if adata.obsm[basis] does not exist. TODO: do we rly need that parameter? Would be fine to always expect a pre calculated embedding!

  • umap_params (dict) – Parameters for sc.tl.umap(). Only applicable if adata.obsm[basis] does not exist. TODO: do we rly need that parameter? Would be fine to always expect a pre calculated embedding!

  • kwargs – Keyword arguments of selection_histogram().

Example

(Takes a few minutes to calculate)

import spapros as sp
adata = sp.ut.get_processed_pbmc_data()
selector = sp.se.ProbesetSelector(adata, "celltype", n=30, verbosity=0)
selector.select_probeset()

selector.plot_clf_genes(n_cols=4,celltypes=["FCGR3A+ Monocytes","Megakaryocytes"])
../_images/Selector_plot_clf_genes.png

TODO: this function and pl.clf_genes_umaps need to be tested on all argument combinations + can be optimized.