API Reference#

Core workflow#

`multi_sample_nmf`	Run NMF independently for each sample and collect gene loadings.
`derive_nmf_metaprograms`	Derive consensus meta-programs (MPs) across all NMF programs.
`get_metaprogram_metrics`	Compute per-meta-program quality metrics.
`compute_mp_scores`	Compute module scores for each meta-program.
`find_optimal_n_mp`	Determine the optimal number of meta-programs via silhouette score.

Filtering and editing#

`filter_by_scores`	Identify meta-programs that fall below quality-metric thresholds.
`filter_for_significant_obs`	Keep only MPs whose scores differ significantly between one condition and the rest.
`drop_meta_programs`	Remove one or more meta-programs from result_dict in-place.
`merge_meta_programs`	Merge two or more meta-programs into a single consensus program.
`propose_mp_actions`	Suggest DROP or MERGE actions for low-quality meta-programs.
`apply_mp_actions`	Apply DROP and MERGE actions proposed by `propose_mp_actions()`.

GSEA#

`run_gsea_pipeline`	Run the full GSEA workflow for all meta-programs.
`plot_gsea_results`	Bar plots of enriched terms per meta-program.

Visualisation#

`plot_heatmap`	Plot a cosine-similarity heatmap with optional cluster boundaries.
`plot_convergence_plots`	Plot NMF convergence curves for one or multiple samples.
`plot_mp_scores_on_umap`	Overlay MP module scores on a UMAP embedding.
`show_distribution_over_obs`	Violin plots of MP module scores across obs categories.
`recompute_pca_umap`	Recompute PCA and UMAP on adata in-place.

Gene-name translation#

translate_gene_names

Translate gene names in results_mp using a mapping from adata.var.

Persistence#

`save`	Save a kaleidocell result object to disk.
`load`	Load a kaleidocell result object from disk.

Report#

Generate a self-contained tabbed HTML report of kaleidocell results.

Bundled files#

Registry of data files bundled with kaleidocell.