API

Models

CauTrigger1L(adata[, n_latent, n_causal, ...])	First-layer causal generative model for perturbation-response modelling.
CauTrigger2L(adata[, n_latent, n_causal, ...])	Second-layer hierarchical causal decomposition model.
CauTrigger3L(adata[, n_latent, n_causal, ...])	Third-layer hierarchical causal decomposition model.

Causal Effect

joint_uncond_v2(params, model, data, index)	Estimate the negative unconditional mutual information -I(α; Ŷ) via Monte Carlo sampling.
joint_uncond_single_dim_v2(params, model, ...)	Estimate -I(z_i; Ŷ) for a single latent dimension i via conditional Monte Carlo sampling.

Utils

pert_plot_vector_field(adata_TF, adata_down, ...)	Perform in silico perturbation and visualize the resulting vector field to show cell state transitions.
pert_plot_stream(adata_TF, adata_down, ...)	Perform in silico perturbation and visualize the resulting streamlines (trajectories) to show cell state transitions.