API reference¶

This page is generated from the docstrings of every public symbol in cycombinepy. Click through to each function for the full parameter list, return type, and source link.

Preprocessing¶

`transform_asinh`	Asinh-transform marker columns of `adata` with an optional derandomization step.
`normalize`	Batch-wise normalize marker columns of `adata`.

Clustering¶

create_som

Train a FlowSOM on the marker columns of adata and store cluster labels.

Correction¶

`batch_correct`	Full cyCombine pipeline: normalize → SOM → per-cluster ComBat.
`correct_data`	Per-cluster ComBat batch correction.

Evaluation¶

`compute_emd`	Per (cluster, marker, batch-pair) 1-D Earth Mover's distance.
`evaluate_emd`	Join uncorrected vs corrected EMD and compute percent reduction.
`compute_mad`	Per (cluster, marker, batch) Median Absolute Deviation.
`evaluate_mad`	Join uncorrected vs corrected MAD and compute percent reduction.

Detection¶

`detect_batch_effect`	Comprehensive batch-effect diagnostic: express + UMAP + MAD summary.
`detect_batch_effect_express`	Quick 3-panel batch-effect summary (EMD heatmap, density, MDS).

Utilities¶

`get_markers`	Return var_names that are not in the non-markers blacklist (case-insensitive).
`check_confound`	Return True if `batch` is confounded with `mod`.

I/O¶

read_fcs_dir

Read all FCS files in data_dir into a single AnnData.

Plotting¶

`plot_density`	Per-marker density plot colored by batch.
`plot_dimred`	Thin wrapper around `scanpy.pl.umap` / `scanpy.pl.pca`.
`plot_emd_heatmap`	Heatmap of mean EMD per (cluster, marker) from `compute_emd()` output.

Advanced evaluation¶

scib_metrics

Run a minimal scib-metrics benchmark on an AnnData.