Plot terminal states

This example shows how to compute and plot the terminal states of the cell-state transition.

CellRank can be applied to any cell-state transition, be it differentiation, regeneration, reprogramming or other.

import cellrank as cr

adata = cr.datasets.pancreas_preprocessed("../example.h5ad")
adata

Out:

AnnData object with n_obs × n_vars = 2531 × 2000
    obs: 'day', 'proliferation', 'G2M_score', 'S_score', 'phase', 'clusters_coarse', 'clusters', 'clusters_fine', 'louvain_Alpha', 'louvain_Beta', 'initial_size_unspliced', 'initial_size_spliced', 'initial_size', 'n_counts', 'velocity_self_transition', 'dpt_pseudotime'
    var: 'highly_variable_genes', 'gene_count_corr', 'means', 'dispersions', 'dispersions_norm', 'fit_r2', 'fit_alpha', 'fit_beta', 'fit_gamma', 'fit_t_', 'fit_scaling', 'fit_std_u', 'fit_std_s', 'fit_likelihood', 'fit_u0', 'fit_s0', 'fit_pval_steady', 'fit_steady_u', 'fit_steady_s', 'fit_variance', 'fit_alignment_scaling', 'velocity_genes'
    uns: 'clusters_colors', 'clusters_fine_colors', 'diffmap_evals', 'iroot', 'louvain_Alpha_colors', 'louvain_Beta_colors', 'neighbors', 'pca', 'recover_dynamics', 'velocity_graph', 'velocity_graph_neg', 'velocity_params'
    obsm: 'X_diffmap', 'X_pca', 'X_umap', 'velocity_umap'
    varm: 'PCs', 'loss'
    layers: 'Ms', 'Mu', 'fit_t', 'fit_tau', 'fit_tau_', 'spliced', 'unspliced', 'velocity', 'velocity_u'
    obsp: 'connectivities', 'distances'

First, we compute the terminal states. By default, we’re using the cellrank.tl.estimators.GPCCA estimator. The parameter cluster_key tries to associate the names of the terminal states with cluster labels, whereas n_cells controls how many cells we take from each terminal state as categorical observation - this is only available to the above mentioned estimator. We can show some plots of interest by specifying show_plots=True.

cr.tl.terminal_states(
    adata,
    cluster_key="clusters",
    n_cells=30,
    softmax_scale=4,
    n_states=3,
    show_progress_bar=False,
)

Out:

INFO: Using pre-computed schur decomposition

We can now plot the terminal states. By default, when using cellrank.tl.estimators.GPCCA, we plot continuous membership vectors to visualize individual cells associations with an terminal state.

cr.pl.terminal_states(adata)

We can also plot membership vectors for different terminal states separately state using same_plot=False.

cr.pl.terminal_states(adata, same_plot=False)

Lastly, we can discretize the assignment of cells to terminal states by showing the cells most likely to belong to the terminal state by specifying the discrete parameter.

cr.pl.terminal_states(adata, discrete=True)

To see how to compute and plot the terminal states or the lineages, see Plot initial states or Plot lineages, respectively.

Estimated memory usage: 597 MB