pseudodynamics.plotting_fns.param_plot.params_in_umap

pseudodynamics.plotting_fns.param_plot.params_in_umap(adata, prediction, timepoints=None, param='u', copy=True, cell_of_t=True, log=False, clipping=None, subplot_kws=None, umap_kws=None)[source]

Visaulize the fitted behavior params in umap and by time

Parameters:

  • adata (AnnData)

  • prediction ([tensor, ndarray] , the prediction of shape (n_timepoints, n_cells))

  • timepoints (list of real-time)

  • param (str, the param to visaulize, must be one of ['u', 'g', 'v', 'D'])

  • copy (bool, default to True, will save the params to adata.obs if copy is set to False)

  • cell_of_t (bool, default to True, only visualize cells of each timepoints.) – If set to False, all cells will be shown in each panels.

Returns:

fig, axs

Example

>>> param = 'g'
>>> u_pred = Model.predict_param(DataSet=train_DS_t5, param=param)
>>> adata = train_DS_t5.adata
>>> params_in_umap(adata, u_pred, param=param)