campmc.mt.quality

Contents

campmc.mt.quality#

campmc.mt.quality(adata, *, partition_key='camp', label_key=None, use_rep='X_pca')[source]#

Compute per-metacell quality metrics.

Metrics include compactness and separation on the diffusion map, an INV (intra-metacell variability) score, and optionally purity / label entropy when label_key is provided.

Parameters:
adata AnnData

Annotated data with metacell labels in obs[partition_key].

partition_key str (default: 'camp')

Column in adata.obs holding metacell assignments.

label_key str | None (default: None)

Optional cell-type (or other) labels for purity and entropy.

use_rep str (default: 'X_pca')

Representation used to compute neighbors / diffusion map when X_diffmap is not already present.

Return type:

DataFrame

Returns:

DataFrame One row per metacell with columns such as size, compactness, separation, sc_ratio, INV, and optionally purity / label_entropy.

Examples

import campmc as cp
import scanpy as sc
adata = sc.datasets.pbmc68k_reduced()
cp.partition(adata, method="camp3", gamma=50, random_state=0)
qc = cp.mt.quality(adata, partition_key="camp", label_key="louvain")
qc[["metacell_id", "size", "purity", "sc_ratio"]].head()
metacell_id size purity sc_ratio
0 seed0-50-__all__ 65 0.984615 4.415655
1 seed1-50-__all__ 34 0.911765 10.381738
2 seed10-50-__all__ 3 0.666667 2.374693
3 seed11-50-__all__ 158 0.708861 2.823893
4 seed12-50-__all__ 42 0.809524 1.641697
Contents