combatlearn 2.0.0

Batch-effect harmonization for machine learning frameworks.

combatlearn

combatlearn makes the popular ComBat (and CovBat) batch-effect correction algorithm available for use into machine learning frameworks. It lets you harmonise high-dimensional data inside a scikit-learn Pipeline, so that cross-validation and grid-search automatically take batch structure into account, without data leakage.

Three methods:

• method="johnson" - classic ComBat (Johnson et al., 2007)
• method="fortin" - neuroComBat (Fortin et al., 2018)
• method="chen" - CovBat (Chen et al., 2022)

Installation

pip install combatlearn

Documentation

Full documentation is available at combatlearn.readthedocs.io

The documentation includes:

• Methods Guide
• API Reference

Quick start

import pandas as pd
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import LogisticRegression
from combatlearn import ComBat

df = pd.read_csv("data.csv", index_col=0)
X, y = df.drop(columns="y"), df["y"]

batch = pd.read_csv("batch.csv", index_col=0).squeeze("columns")
diag = pd.read_csv("diagnosis.csv", index_col=0) # categorical
age = pd.read_csv("age.csv", index_col=0) # continuous

pipe = Pipeline([
    ("combat", ComBat(
        batch=batch,
        discrete_covariates=diag,
        continuous_covariates=age,
        method="fortin", # or "johnson" or "chen"
        parametric=True
    )),
    ("scaler", StandardScaler()),
    ("clf", LogisticRegression())
])

param_grid = {
    "combat__mean_only": [True, False],
    "clf__C": [0.01, 0.1, 1, 10],
}

grid = GridSearchCV(
    estimator=pipe,
    param_grid=param_grid,
    cv=5,
    scoring="roc_auc",
)

grid.fit(X, y)

print("Best parameters:", grid.best_params_)
print(f"Best CV AUROC: {grid.best_score_:.3f}")

For a full example of how to use combatlearn see the notebook demo

ComBat parameters

The following section provides a detailed explanation of all parameters available in the scikit-learn-compatible ComBat class. For complete API documentation, see the API Reference.

Main Parameters

Parameter	Type	Default	Description
batch	array-like or pd.Series	required	Vector indicating batch assignment for each sample. This is used to estimate and remove batch effects.
discrete_covariates	array-like, pd.Series, or pd.DataFrame	None	Optional categorical covariates (e.g., sex, site). Only used in "fortin" and "chen" methods.
continuous_covariates	array-like, pd.Series or pd.DataFrame	None	Optional continuous covariates (e.g., age). Only used in "fortin" and "chen" methods.

Algorithm Options

Parameter	Type	Default	Description
method	str	"johnson"	ComBat method to use: "johnson" - Classical ComBat (Johnson et al. 2007) "fortin" - ComBat with covariates (Fortin et al. 2018) "chen" - CovBat, PCA-based correction (Chen et al. 2022)
parametric	bool	True	Whether to use the parametric empirical Bayes formulation. If False, a non-parametric iterative scheme is used.
mean_only	bool	False	If True, only the mean is corrected, while variances are left unchanged. Useful for preserving variance structure in the data.
reference_batch	str or None	None	If specified, acts as a reference batch - other batches will be corrected to match this one.
covbat_cov_thresh	float, int	0.9	For "chen" method only: Cumulative variance threshold $]0,1[$ to retain PCs in PCA space (e.g., 0.9 = retain 90% explained variance). If an integer is provided, it represents the number of principal components to use.
eps	float	1e-8	Small jitter value added to variances to prevent divide-by-zero errors during standardization.

Batch Effect Correction Visualization

The plot_transformation method allows to visualize the ComBat transformation effect using dimensionality reduction, showing the before/after comparison of data transformed by ComBat using PCA, t-SNE, or UMAP to reduce dimensions for visualization.

For further details see the API Reference and the notebook demo.

Batch Effect Metrics

The compute_batch_metrics method provides quantitative assessment of batch correction quality. It computes metrics including Silhouette coefficient, Davies-Bouldin index, kBET, LISI, and variance ratio for batch effect quantification, as well as k-NN preservation and distance correlation for structure preservation.

For further details see the API Reference and the notebook demo.

Contributing

Pull requests, bug reports, and feature ideas are welcome: feel free to open a PR!

Author

Ettore Rocchi @ University of Bologna

Google Scholar | Scopus

Citation

If combatlearn is useful in your research, please cite the paper introducing this Python package:

Rocchi, E., Nicitra, E., Calvo, M. et al. Combining mass spectrometry and machine learning models for predicting Klebsiella pneumoniae antimicrobial resistance: a multicenter experience from clinical isolates in Italy. BMC Microbiol (2026). https://doi.org/10.1186/s12866-025-04657-2

@article{Rocchi2026,
  author    = {Rocchi, Ettore and Nicitra, Emanuele and Calvo, Maddalena and Cento, Valeria and Peiretti, Laura and Asif, Zian and Menchinelli, Giulia and Posteraro, Brunella and Sala, Claudia and Colosimo, Claudia and Cricca, Monica and Sambri, Vittorio and Sanguinetti, Maurizio and Castellani, Gastone and Stefani, Stefania},
  title     = {Combining mass spectrometry and machine learning models for predicting Klebsiella pneumoniae antimicrobial resistance: a multicenter experience from clinical isolates in Italy},
  journal   = {BMC Microbiology},
  year      = {2026},
  doi       = {10.1186/s12866-025-04657-2},
  url       = {https://doi.org/10.1186/s12866-025-04657-2}
}

Acknowledgements

This project builds on the excellent work of the ComBat family of harmonisation methods. Please consider citing the original papers:

• ComBat - Johnson WE, Li C, Rabinovic A. Biostatistics. 2007. doi: 10.1093/biostatistics/kxj037

• neuroCombat - Fortin JP et al. Neuroimage. 2018. doi: 10.1016/j.neuroimage.2017.11.024

• CovBat - Chen AA et al. Hum Brain Mapp. 2022. doi: 10.1002/hbm.25688