Velocity Entrain Analysis

Python function takes in a dataset containing velocities and predicts ligands that may be driving the observed velocities. Does not consider spatial proximity. For spatial data, use `get_velocity_ligands_spatial``

Usage

get_velocity_ligands(
  adata,
  velocity_cluster_key = "velocity_clusters",
  n_top_genes = None,
  vclusters = None,
  sender_clusters = None,
  sender_cluster_key = None,
  ligand_target_matrix = None,
  expression_proportion_cutoff = 0.05,
  organism = "human",
  n_jobs = 4
)

Arguments

adata

An anndata object containing velocities calculated with scvelo

velocity_cluster_key

Column name in adata.obs denoting velocities clustered by entrain.cluster_velocities()

n_top_genes

Default None. Number of top velocity genes to consider for each cluster for scvelo.tl.recover_dynamics().

vclusters

Default None. If you would like to limit the analysis to a set of velocity clusters, you can define them here as an array.

sender_clusters

If you want to specify a population that constitutes the niche that interacts with your differentiating cells, you can specify the cell type annotations here. For example ["Stromal Cells"]

sender_cluster_key

Column name in adata.obs that denotes the sender_clusters. For example "Celltype".

ligand_target_matrix

NicheNet database

expression_proportion_cutoff

Default 0.05. Population fraction threshold for including a ligand or receptor for Entrain analysis. A ligand or receptor is only included in analysis if it is expressed above this proportion. Specifically, if this is set to 0.05, a ligand or receptor is included in analysis if it has > 0 counts in 5% of cells in adata.X, where adata is subset to the cells belonging to the velocity cluster (for receptors); or spots belonging to the immediate spatial environment of the velocity cluster (for ligands).

organism

Default "human". Either one of "human" or "mouse". This argument feeds into omnipath-pypath module to convert human genes into mouse genes, if "mouse" is specified.

n_jobs

Default 4. Number of parallel jobs to use for scvelo.tl.recover_dynamics()

Value

An anndata object with Entrain results in anndata.uns