# Batch effects and harmonisation

Source: https://onco.cc/terms/batch-effects/  
OnCo record `batch-effects` (Term). Data CC BY-NC 4.0, attribute "Data from OnCo (onco.cc)"; commercial use needs a licence.

## TL;DR

A batch effect is a difference in the data caused by when, where or how samples were processed rather than by biology; if it lines up with the outcome, a model learns the batch instead of the disease.

## Summary

In molecular biology a batch effect occurs when non-biological factors cause changes in the data, and it leads to wrong conclusions when its causes correlate with the outcome of interest (Wikipedia). Moving between RNA-seq and microarray, or between TCGA and a hospital cohort, is the extreme case. ComBat, from Johnson, Li and Rabinovic, is the standard empirical Bayes correction; per-gene z-scoring, quantile normalisation and rank transforms are the alternatives that need no batch labels. Any correction fitted on test data leaks.

## Fields

- Kind: Term
- Last checked: 2026-09-24
- Also known as: batch effects; batch effect; batch correction; ComBat; platform effect; cross-platform harmonisation; harmonisation across platforms
- Tags: cansim-terms

## Notes

- Listed in the CanSim terms map 1.0.0 (docs/onco/terms.json, generated 2026-09-24), CC BY 4.0, attribution: CanSim project, an open, public-data-first cancer foundation-model programme; CanSim page path /terms/batch-effects.

## Sources

- Wikipedia: https://en.wikipedia.org/wiki/Batch_effect
- Johnson, Li and Rabinovic, Adjusting batch effects in microarray expression data using empirical Bayes methods (ComBat, Biostatistics 2007): https://doi.org/10.1093/biostatistics/kxj037
- sva Bioconductor package (ComBat): https://bioconductor.org/packages/release/bioc/html/sva.html
- Wikipedia: https://en.wikipedia.org/wiki/Batch_effect

## Connected records

- terms: [Cancer AI vocabulary (CanSim terms map)](https://onco.cc/terms/cancer-ai-vocabulary/), [Data leakage in model evaluation](https://onco.cc/terms/data-leakage/), [Domain shift and domain adaptation (cell line to patient)](https://onco.cc/terms/domain-adaptation/), [Magnification (20x, 40x) and microns per pixel](https://onco.cc/terms/magnification/), [Microarray expression data](https://onco.cc/terms/microarray-expression/), [Quantile normalisation, rank transforms and z-scores](https://onco.cc/terms/quantile-normalisation/), [TPM, FPKM and raw counts (expression units)](https://onco.cc/terms/tpm-fpkm-counts/)

---
JSON: https://onco.cc/api/v1/entities/batch-effects.json