Methods that remove batch effects while retaining group differences may lead to exaggerated confidence in downstream analyses

By Vegard Nygaard, E. Rødland, Eivind Hovig. Biostatistics, 2015

325 citations recorded by the source corpus.

Abstract Removal of, or adjustment for, batch effects or center differences is generally required when such effects are present in data. In particular, when preparing microarray gene expression data from multiple cohorts, array platforms, or batches for later analyses, batch effects can have confounding effects, inducing spurious differences between study groups. Many methods and tools exist for removing batch effects from data. However, when study groups are not evenly distributed across batches, actual group differences may induce apparent batch differences, in which case batch adjustments may bias, usually deflate, group differences. Some tools therefore have the option of preserving the difference between study groups, e.g. using a two-way ANOVA model to simultaneously estimate both group and batch effects. Unfortunately, this approach may systematically induce incorrect group differences in downstream analyses when groups are distributed between the batches in an unbalanced manner. The scientific community seems to be largely unaware of how this approach may lead to false discoveries.

Subfields: Molecular Biology, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/9aa2efe400c62b2505f792e8b26840c18fd9499e

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Molecular Biology

1. Aldons J. Lusis
2. Yuval Kluger
3. Yibin Wang

In Statistics and Probability

1. Uri Shaham
2. Yuval Kluger

Citations in the corpus

• WILDS: A Benchmark of in-the-Wild Distribution Shifts (2020, unknown): “These batch effects make it difficult to draw conclusions from data across experimental batches (Leek et al., 2010; Parker & Leek, 2012; Soneson et al., 2014; Nygaard et al., 2016; Caicedo et al., 2017).”
• Systems Genetics Approach Identifies Gene Pathways and Adamts2 as Drivers of Isoproterenol-Induced Cardiac Hypertrophy and Cardiomyopathy in Mice (2016, background): “Thus, we do not believe that our data suffer from the potential batch artifacts as reported in Nygaard and Rødland, 2016.”
• Extracting Biological Insights from the Project Achilles Genome-Scale CRISPR Screens in Cancer Cell Lines (2019, background): “First, corrections that utilize phenotype have the potential to introduce misleading results ​ 14 .”
• Probabilistic harmonization and annotation of single‐cell transcriptomics data with deep generative models (2019, unknown): “Since the problem of data harmonization of single-cell transcrip-tomics is difficult and can potentially lead to over-correction (Appendix Fig S1; Nygaard et al , 2016), we propose a fully generative method as a robust and principled approach to address it.”
• Removal of batch effects using distribution‐matching residual networks (2016, background): “The simplest and most common (Nygaard et al., 2016) adjustment is zero centering, i.”