Integration
mina.integration.patpy.PrecomputedSampleRepresentation
Bases: SampleRepresentationMethod
Wrap a precomputed sample-by-feature matrix for patpy tools.
This is useful when sample-level representations were computed outside patpy, for example using MINA, MOFA, PCA, or another latent-variable model.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
representation
|
DataFrame
|
DataFrame with samples in rows and latent variables/features in columns. |
required |
metadata
|
DataFrame or None
|
Optional sample-level metadata. Its index must contain the same sample
IDs as |
None
|
metric
|
str
|
Distance metric passed to |
'euclidean'
|
seed
|
int
|
Random seed used by patpy embedding methods. |
67
|
prepare_anndata(adata=None)
Skip AnnData preparation.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
adata
|
AnnData or None
|
Ignored placeholder for compatibility with the patpy interface. |
None
|
Returns:
| Type | Description |
|---|---|
None
|
The representation is already sample-level and no preparation is needed. |
calculate_distance_matrix(force: bool = False, dist: str | None = None) -> np.ndarray
Calculate a sample-by-sample distance matrix.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
force
|
bool
|
If True, recompute distances even when cached distances exist. |
False
|
dist
|
str or None
|
Distance metric passed to |
None
|
Returns:
| Name | Type | Description |
|---|---|---|
distances |
ndarray
|
Square sample-by-sample distance matrix. |
to_adata(metadata: pd.DataFrame | None = None, *args, **kwargs) -> ad.AnnData
Convert the precomputed sample representation to an AnnData object.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
metadata
|
DataFrame or None
|
Optional metadata to pass to the patpy conversion. If None, the metadata stored on the representation is used. |
None
|
*args
|
tuple
|
Additional positional arguments passed to the parent method. |
()
|
**kwargs
|
dict
|
Additional keyword arguments passed to the parent method. |
{}
|
Returns:
| Name | Type | Description |
|---|---|---|
adata |
AnnData
|
AnnData representation produced by the parent patpy method. |