regvelo.tools.markov_density_simulation

regvelo.tools.markov_density_simulation(adata, T, start_indices, terminal_indices, terminal_states, n_steps=100, n_simulations=1000, method='stepwise', seed=0)[source]

Simulate transitions on a velocity-derived Markov transition matrix.

Parameters:

  • adata (AnnData) – Annotated data object.

  • T (ndarray) – Transition matrix of shape (n_cells, n_cells).

  • start_indices (list[int]) – Indices of starting cells.

  • terminal_indices (list[int]) – Indices of terminal cells.

  • terminal_states (list[str]) – Labels of terminal states corresponding to cells in adata.obs["term_states_fwd"].

  • n_steps (int) – Maximum number of steps per simulation.

  • n_simulations (int) – Number of simulations per starting cell.

  • method (Literal['stepwise', 'one-step']) –

    Simulation method:

    • 'stepwise': simulate trajectories step by step.

    • 'one-step': sample directly from T^n.

  • seed (int) – Random seed for reproducibility.

Return type:

int

Returns:

: Total number of simulations performed. Also updates adata with the following fields:

  • adata.obs["visits"]: Number of visits to each terminal cell.

  • adata.obs["visits_dens"]: Density of visits to each terminal cell, calculated as the proportion of visits relative to the total number of simulations.