cellrank.external.kernels.WOTKernel.compute_initial_growth_rates

WOTKernel.compute_initial_growth_rates(proliferation_key=None, apoptosis_key=None, organism=None, beta_min=0.3, beta_max=1.7, delta_min=0.3, delta_max=1.7, key_added=None, **kwargs)[source]

Estimate initial growth rates using a birth-death process as described in [Schiebinger et al., 2019].

The doubling time is defined as \(\frac{\ln 2}{\beta - \delta}\) (similarly defined for half-time). The logistic function is used to transform the birth/death rate scores and to smoothly interpolate between specified minimum and maximum birth/death rates.

Parameters
  • proliferation_key (Optional[str]) – Key in anndata.AnnData.obs where the birth rate score is saved.

  • apoptosis_key (Optional[str]) – Key in anndata.AnnData.obs where the death rate score is saved.

  • organism (Optional[Literal[‘human’, ‘mouse’]]) – Organism for which to calculate the birth/death scores, if they cannot be found in adata. In this case, scanpy.tl.score_genes() is used to calculate the scores based on an organism-dependent set of marker genes for proliferation and apoptosis.

  • beta_min (float) – Minimum birth rate.

  • beta_max (float) – Maximum birth rate.

  • delta_min (float) – Minimum death rate.

  • delta_max (float) – Maximum death rate.

  • key_added (Optional[str]) – Key in adata .obs where to add the estimated growth rates. If None, just return them.

  • kwargs (Any) – Keyword arguments for scanpy.tl.score_genes(). Only used when proliferation_key or apoptosis_key cannot be found in adata.AnnData.obs.

Return type

Optional[Series]

Returns

pandas.Series The estimated initial growth rates if key_added = None, otherwise None.

Notes

If you don’t have access to proliferation/apoptosis gene sets, you can use the ones defined in cellrank for a specific organism. Alternatively, you can also use WOT without an estimate of initial growth rates. In that case, make sure to use several iterations in cellrank.external.kernels.WOTKernel.compute_transition_matrix() by increasing the growth_iters parameter. A value around 3 works well in most cases.

The markers used here were taken from the following sources:

For more information about WOT, see the official tutorial.