regvelo.tools.in_silico_block_simulation

regvelo.tools.in_silico_block_simulation(model, adata, TF, effects=0, cutoff=0.001, customized_GRN=None, batch_size=None)[source]

Perform an in silico transcription factor (TF) regulon knock-out by modifying the gene regulatory network (GRN) in a trained RegVelo model.

Parameters:

  • model (str) – Path to the saved RegVelo model.

  • adata (AnnData) – Annotated data matrix.

  • TF (Union[str, list[str]]) – Transcription factor(s) to be knocked out (their regulons will be silenced).

  • effect – Coefficient used to replace weights in GRN

  • cutoff (float) – Threshold to determine which links in the GRN are considered active and should be muted.

  • customized_GRN (Optional[Tensor]) – A custom perturbed GRN weight matrix to directly replace the original GRN.

  • batch_size (Optional[int]) – the batch size used for velocity and latent time generation.

  • effects (float)

Return type:

tuple[AnnData, REGVELOVI]

Returns:

:

  • Perturbed annotated data object with RegVelo outputs.

  • RegVelo model with modified GRN.