A cautionary tale of batch corrections on confounded microbiome community profiles
We use a case study of seed microbiomes to assess the performance of five batch effects correction algorithms (BECAs) (zero-mean centering (ZMC), Ratio-A, conditional quantile regression (ConQuR), partial least squares discriminant analysis (PLSDA), and weighted PLSDA on a confounded data set in which the covariate of interest (plant species) does not appear in all batches (studies). We show incomplete batch effects removal in all cases, which calls for careful application of and future work on BECAs.
Authors
- Alicia J. Foxx (ORCID: https://orcid.org/0000-0001-5504-2986)
- Adam R. Rivers (ORCID: https://orcid.org/0000-0002-3703-834X)
Institutions
- Northwestern University (US)
- Agricultural Research Service (US)
- United States Department of Agriculture (US)
Publication Details
- Journal
- iMetaOmics.
- Published
- 2025-05-19
- DOI
- https://doi.org/10.1002/imo2.70025
- Citations
- 1
- Primary Topic
- Gut microbiota and health
- Type
- article
- Field-Weighted Citation Impact
- 0.81
Funders
- U.S. Department of Energy
- U.S. Department of Agriculture
- Oak Ridge Associated Universities
- Oak Ridge Institute for Science and Education
- Agricultural Research Service