stt.tl submodule

Module contents (Calling by stt.tl.function_name)

stt.tl._dynamical_analysis.dynamical_iteration(adata, n_states=None, n_states_seq=None, n_iter=10, return_aggr_obj=True, weight_connectivities=0.2, n_components=20, n_neighbors=100, thresh_ms_gene=0, thresh_entropy=0.1, use_spatial=False, spa_weight=0.5, spa_conn_key='spatial', monitor_mode=False, l2=0.1)

Perform dynamical iteration on the given AnnData object. The function updates the attractor states, the tensor, the averaged velocity, and the entropy at each iteration.

Parameters:

adata: AnnData object

Annotated data matrix.

n_states: int, optional (default: None)

Number of attractor states.

n_states_seq: list, optional (default: None)

List of number of attractor states for each iteration.

n_iter: int, optional (default: 10)

Number of iterations.

return_aggr_obj: bool, optional (default: False)

Whether to return the aggregated object.

weight_connectivities: float, optional (default: 0.2)

Weight of connectivities.

n_components: int, optional (default: 20)

Number of components.

n_neighbors: int, optional (default: 100)

Number of neighbors.

thresh_ms_gene: int, optional (default: 0)

Threshold for mean spliced gene expression.

thresh_entropy: float, optional (default: 0.1)

Threshold for entropy.

use_spatial: bool, optional (default: False)

Whether to use spatial information.

spa_weight: float, optional (default: 0.5)

Weight of spatial information.

spa_conn_key: str, optional (default: ‘spatial’)

Key for spatial connectivities.

stop_cr: str, optional (default: ‘abs’)

Stopping criterion for iteration.

monitor_mode: bool, optional (default: False)

Whether to use monitor mode.

l2: float, optional (default: 0.1)

Regularization parameter in tensor estimation.

Returns:

Default none, but updates the adata.uns with the following keys: da_out: dict

Dictionary of the results of dynamical analysis.

gene_subset: list

List of selected multi-stable genes.

entropy: np.ndarray

Array of entropy values.

speed: np.ndarray

Array of speed values.

attractor: np.ndarray

Array of attractor states.

tensor_v_aver: np.ndarray

Array of averaged tensor by attractor membership of cells.

If return_aggr_obj is True, the aggregated object with both spliced and unspliced counts of multi-stable genes is returned.

stt.tl._construct_landscape.construct_landscape(sc_object, thresh_cal_cov=0.3, scale_axis=1.0, scale_land=1.1, N_grid=100, coord_key='X_umap')

Function to construct the landscape of the multi-stable attractors

Parameters

  • sc_object (AnnData object) – Single cell data object

  • thresh_cal_cov (float) – Threshold to calculate the covariance matrix

  • scale_axis (float) – Scaling factor for the axis

  • scale_land (float) – Scaling factor for the landscape

  • N_grid (int) – Number of grid points for the landscape

  • coord_key (str) – Key of the coordinates in the sc_object.obsm

Returns

  • None, but updates the sc_object.uns with the following keys

  • land_out (dict) – Dictionary of landscape values and grid points

stt.tl._pathway_analysis.compute_pathway(adata, adata_aggr, db_name, gene_num=3)

Compute tensor similarities among pathways

Parameters

  • adata (AnnData) – Annotated data matrix

  • adata_aggr (AnnData) – Aggregated data matrix

  • db_name (str) – Name of the database

  • gene_num (int) – Minimum number of genes in the pathway overlapped with STT multi-stable genes

Returns

  • None, but updates adata.uns with the following

  • pathway_select (dict) – Selected pathways satisfying the gene_num condition

  • pathway_embedding (np.ndarray) – UMAP embedding of the pathway similarities

  • pathway_labels (np.ndarray) – Cluster labels of the pathway embedding