Tools (sdq.tl)#

sdq.tl.annotate_cell_state
scdiffeq.tools._annotate_cell_state.annotate_cell_state(adata_sim: AnnData, kNN: kNN, obs_key: str = 'state', use_key: str = 'X', silent: bool = False) → None[source]

Use a kNN Graph to annotate simulated cell states.

Parameters:

  • adata_sim (AnnData) – Simulated data object in the format of anndata.AnnData, the (annotated) single-cell data matrix of shape n_obs × n_vars. Rows correspond to cells and columns to genes. For more: [1](https://anndata.readthedocs.io/en/latest/).

  • kNN (kNN) – k-nearest neighbor graph.

  • obs_key (str, optional) – Observation key. Default is “state”.

  • use_key (str, optional) – Key to use for the basis. Default is “X”.

  • silent (bool, optional) – If True, suppresses informational messages. Default is False.

Return type:

None

References

More: sdq.tl.annotate_cell_state

sdq.tl.annotate_cell_fate
scdiffeq.tools._annotate_cell_fate.annotate_cell_fate(adata_sim: AnnData, state_key: str = 'state', key_added: str = 'fate', time_key: str = 't', sim_key: str = 'sim', silent: bool = False, *args, **kwargs) → None[source]

Annotate the fate of a simulated cell trajectory, given annotated states.

Parameters:

adata_sim (AnnData) –

Simulated AnnData object.

state_key (str)

Key in adata_sim.obs corresponding to annotated states.

key_added (str)

Key added to adata_sim.obs to indicate the fate of each trajectory. Default: “fate”

time_key (str)

Key in adata_sim.obs corresponding to time.

sim_key (str)

Key in adata_sim.obs corresponding to each individual simulation.

silent (bool)

If True, suppresses informational messages.

Return type:

None

More: sdq.tl.annotate_cell_fate

sdq.tl.kNN
scdiffeq.tools._knn.kNN(adata: AnnData, use_key: str = 'X_pca', n_neighbors: int = 20, metric: str = 'euclidean', space: Space | None = None)[source]

k-Nearest Neighbors container using voyager backend.

This class provides a kNN graph interface for single-cell data stored in AnnData objects. It uses voyager (Spotify’s HNSW implementation) for efficient approximate nearest neighbor search.

Parameters:

  • adata – AnnData object containing the data.

  • use_key – Key to fetch data from adata (e.g., “X_pca”). Default: “X_pca”.

  • n_neighbors – Number of neighbors to return in queries. Default: 20.

  • metric – Distance metric to use. One of “euclidean”, “cosine”, “inner_product”. Default: “euclidean”.

  • space – Alternative to metric - directly specify voyager.Space. If provided, overrides metric parameter.

scdiffeq.tools._knn.adata

The AnnData object.

scdiffeq.tools._knn.use_key

Key used to fetch data.

scdiffeq.tools._knn.n_neighbors

Number of neighbors for queries.

scdiffeq.tools._knn.space

The voyager.Space used for distance computation.

More: sdq.tl.kNN

sdq.tl.simulate
scdiffeq.tools._simulation.simulate(adata: AnnData, diffeq: LightningModule, idx: Index | None = None, use_key: str = 'X_pca', time_key: str = 'Time point', N: int | None = 1, t: Tensor | None = None, dt: float | None = 0.1, device: device | None = device(type='cpu'), *args, **kwargs) → AnnData[source]

Simulate trajectories by sampling from an scDiffEq model.

Parameters:

  • adata (AnnData) – Input AnnData object.

  • idx (pd.Index) – Cell indices (corresponding to adata) from which the model should initiate sampled trajectories.

  • diffeq (lightning.LightningModule) – The differential equation model.

  • use_key (str, optional) – adata accession key for the input data. Default is “X_pca”.

  • N (int, optional) – Number of trajectories to sample from the model. Default is 2000.

  • device (device, optional) – Device to run the simulation on. Default is True.

Returns:

AnnData object encapsulating scDiffEq model simulation.

Return type:

AnnData

More: sdq.tl.simulate

sdq.tl.perturb

More: sdq.tl.perturb