Plotting

Methods to compute and visualize velocity graphs and trajectories

Cell2fate_DynamicalModel.compute_velocity_graph_Bergen2020(adata, n_neighbours=None, full_posterior=True, spliced_key='Ms', velocity_key='velocity')

Computes a “velocity graph” similar to the method in: “Bergen et al. (2020), Generalizing RNA velocity to transient cell states through dynamical modeling”

Parameters:

• adata – AnnData object with velocity information in adata.layers['velocity'] (expectation value) or adata.uns['velocity_posterior'] (full posterior). Also normalized spliced counts in adata.layers['spliced_nor']

• n_neighbours – How many nearest neighbours to consider (all non nearest neighbours have edge weights set to 0). If not specified, 10% of the total number of cells is used.

• full_posterior – Whether to use full posterior to compute velocity graph (otherwise expectation value is used).

• velocity_key – Key to access velocity information in adata.

• spliced_key – Key to access normalized spliced counts in adata.

Returns:

Velocity graph

Return type:

Numpy.ndarray

Cell2fate_DynamicalModel.compute_and_plot_module_velocity(adata, delete=True, plot=True, save=None, plotting_kwargs={'color': 'clusters', 'legend_fontsize': 10, 'legend_loc': 'right_margin', 'min_mass': 4})

Computes the RNA velocity produced by each module, as well as associated “velocity graph” and then plots results on a UMAP based on the method in: “Bergen et al. (2020), Generalizing RNA velocity to transient cell states through dynamical modeling”

Parameters:

• adata – AnnData object with spliced and unspliced count data.

• delete – Whether to delete computed layers after processing.

• plot – Whether to plot the results.

• save – Filepath to save the plot.

• plotting_kwargs – Keyword arguments for plotting.

Cell2fate_DynamicalModel.compute_and_plot_total_velocity(adata, delete=True, plot=True, save=None, plotting_kwargs={'color': 'clusters', 'legend_fontsize': 10, 'legend_loc': 'right_margin'}, return_adata=False)

Computes total RNA velocity, as well as associated “velocity graph” and then plots results on a UMAP based on the method in: “Bergen et al. (2020), Generalizing RNA velocity to transient cell states through dynamical modeling”

Parameters:

• adata – AnnData object with spliced and unspliced count data.

• delete – Whether to delete computed layers after processing.

• plot – Whether to plot the results.

• save – Filepath to save the plot.

• plotting_kwargs – Keyword arguments for plotting.

Cell2fate_DynamicalModel.visualize_module_trajectories(adata, chosen_module, delete=True, plot=True, save=None, smooth=None, min_mass=None, n_neighbors=None, cutoff_perc=None, plotting_kwargs={'cmap': 'Greys', 'color': 'clusters', 'dpi': 300, 'legend_fontsize': 10, 'legend_loc': 'on data'})

Visualize relative module activation trajectories using velocity-based embedding.

Parameters:

• adata – AnnData object containing cell information and embeddings.

• chosen_module – The number / name of the chosen module.

• delete – Delete temporary data structures after use, by default True.

• plot – Whether to generate the plot, by default True.

• save – File path to save the generated plot, by default None.

• smooth – Smoothing parameter for grid-based velocity calculations, by default None.

• min_mass – Minimum cell mass for grid-based velocity calculations, by default None.

• n_neighbors – Number of neighbors for grid-based velocity calculations, by default None.

• cutoff_perc – Cutoff percentile for adjusting grid-based velocity calculations, by default None.

• plotting_kwargs – Additional keyword arguments for customizing the plot appearance, by default {"color": 'clusters', 'legend_fontsize': 10, 'legend_loc': 'on data', 'dpi': 300, 'cmap': 'inferno'}.

Cell2fate_DynamicalModel.compute_and_plot_total_velocity_scvelo(adata, delete=True, plot=True, save=None, plotting_kwargs={'color': 'clusters', 'legend_fontsize': 10, 'legend_loc': 'right_margin'})

Computes total RNA velocity, as well as associated “velocity graph” and then plots results on a UMAP based on the method in: “Bergen et al. (2020), Generalizing RNA velocity to transient cell states through dynamical modeling”

Parameters:

• adata – AnnData object with spliced and unspliced count data.

• delete – Whether to delete computed layers after processing.

• plot – Whether to plot the results.

• save – Filepath to save the plot.

• plotting_kwargs – Keyword arguments for plotting.

utils.compute_velocity_graph_Bergen2020(n_neighbours=None, full_posterior=True, spliced_key='Ms')

Computes a “velocity graph” similar to the method in: “Bergen et al. (2020), Generalizing RNA velocity to transient cell states through dynamical modeling”

Parameters:

• adata – anndata object with velocity information in adata.layers['velocity'] (expectation value) or adata.uns['velocity_posterior'] (full posterior). Also normalized spliced counts in adata.layers[‘spliced_norm’].

• n_neighbours – how many nearest neighbours to consider (all non nearest neighbours have edge weights set to 0) if not specified, 10% of the total number of cells is used.

• full_posterior – whether to use full posterior to compute velocity graph (otherwise expectation value is used)

Returns:

Velocity graph

Return type:

Array

utils.plot_velocity_umap_Bergen2020(use_full_posterior=True, n_neighbours=None, plotting_kwargs=None, save=False, spliced_key='Ms')

Visualizes RNAvelocity with arrows on a UMAP, using the method introduced in “Bergen et al. (2020), Generalizing RNA velocity to transient cell states through dynamical modeling” The method computes a “velocity graph” before plotting (see the referenced paper for details), unless such a graph is already available. The graph is based on the full velocity posterior distribution if available, otherwise it is based on the velocity expectation values. Velocity is expected in adata.layers['velocity'] or adata.uns['velocity_posterior'] (for the full posterior) and the graph is saved/expected in adata.layers['velocity_graph'].

Parameters:

• adata – AnnData object with velocity information

• use_full_posterior – Use full posterior to compute velocity graph (if available)

• plotting_kwargs

Methods to plot module summary statistics and activations

Cell2fate_DynamicalModel.compare_module_activation(adata, chosen_modules, time_max=None, time_min=0, save=None, ncol=1)

Compares the activation of chosen modules across time.

Parameters:

• adata – AnnData object.

• chosen_modules – List of module indices to compare.

• time_max – Maximum time point for comparison.

• time_min – Minimum time point for comparison.

• save – Filepath to save the plot.

• ncol – Number of columns in the legend.

Cell2fate_DynamicalModel.plot_module_summary_statistics(adata, save=None)

Plots weight, activation, velocity, switch ON/OFF time histograms for each module.

Parameters:

• adata – AnnData object containing single-cell RNA sequencing data.

• save – File path to save the plot. If not provided, the plot will not be saved.

Other methods to plot top features, genes, and technical variables

Cell2fate_DynamicalModel.plot_top_features(adata, tab, chosen_modules, mode='all genes', n_top_features=3, save=False, process=True)

Plot top features for chosen modules.

Parameters:

• adata – AnnData object.

• tab – Table containing feature rankings.

• chosen_modules – List of module indices to plot.

• mode – Mode for selecting features.

• n_top_features – Number of top features to plot.

• save – Whether to save the plot.

• process – Whether to preprocess the data.

Cell2fate_DynamicalModel.plot_genes(adata, chosen_clusters, marker_genes, cluster_key='clusters', save=None)

Plot expression of marker genes across chosen clusters.

Parameters:

• adata – AnnData object.

• chosen_clusters – List of cluster names to include.

• marker_genes – List of marker genes to plot.

• cluster_key – Key in adata.obs storing cluster information.

• save – Filepath to save the plot.

Cell2fate_DynamicalModel.plot_technical_variables(adata, save=False)

Plot posterior of technical variables in the model.

Parameters:

• adata – AnnData object.

• save – Whether to save the plot.