Surrogate variables and the risk of overfitting in neuropsychiatric gene expression analyses
- 1Icahn School of Medicine at Mount Sinai
Abstract
The reproducibility crisis in biomedical research calls for re-evaluation of how common analytic approaches affect results. One such approach, surrogate variable analysis (SVA), is widely used to adjust for unmeasured drivers of variation in transcriptomic differential expression analyses. We systematically quantified the impact of SVA using synthetic data with known ground truth and fourteen brain RNA-seq and microarray datasets spanning Alzheimer's disease, autism spectrum disorder, bipolar disorder, schizophrenia, and Parkinson's disease. In synthetic data, SVA usage frequently led to high false positive rates. This issue was not revealed by standard diagnostics nor resolved by strategies such as maximizing cross-dataset signal correlation. In real data, using SVA reduced cross-cohort reproducibility of differential expression signatures regardless of the number of surrogate variables included. These findings show that SVA risks inducing overfitting at realistic signal levels, advocating for a cautious, case-by-case application, with a preference for modeling known covariates.
Related
- Conservation of gene co-expression networks and neuropsychiatric-neurological disease gene programs between mouse and human frontal cortex at single-cell resolution
- Genotypic Responses of Rice to Ammonium and Nitrate under Contrasting Water Regimes: Agronomic, Biochemical, and Gene-Expression Analyses
- Bioinformatic and experimental analyses revealed pathogen-derived pathoPEPs candidates predicted to control plant miRNAs expression and infection
- Dynamic changes in clinicopathological variables and gene expression in dogs with immune-mediated hemolytic anemia
- Comparative Transcriptome and Weighted Gene Co-Expression Network Analyses Reveal a Shared Core Response and Quantitatively Divergent Regulatory Modules Between Drought-Tolerant and Drought-Sensitive Foxtail Millet.
BibTeX
@misc{hoang2026surrogate,
title = {{Surrogate variables and the risk of overfitting in neuropsychiatric gene expression analyses}},
author = {D. Hoang and R. Thompson and A. Lund and E. Kritzer and A. Buxbaum-Grice and D. John and G. Nadkarni and N. Golestani and A. W. Charney and N. Beckmann},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.05.756826},
url = {https://www.biorxiv.org/content/10.64898/2026.10.05.756826v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - Surrogate variables and the risk of overfitting in neuropsychiatric gene expression analyses
AU - D. Hoang
AU - R. Thompson
AU - A. Lund
AU - E. Kritzer
AU - A. Buxbaum-Grice
AU - D. John
AU - G. Nadkarni
AU - N. Golestani
AU - A. W. Charney
AU - N. Beckmann
PY - 2026
DA - 2026/10/05
PB - bioRxiv
DO - 10.64898/2026.10.05.756826
UR - https://www.biorxiv.org/content/10.64898/2026.10.05.756826v1
AB - The reproducibility crisis in biomedical research calls for re-evaluation of how common analytic approaches affect results. One such approach, surrogate variable analysis (SVA), is widely used to adjust for unmeasured drivers of variation in transcriptomic differential expression analyses. We systematically quantified the impact of SVA using synthetic data with known ground truth and fourteen brain RNA-seq and microarray datasets spanning Alzheimer's disease, autism spectrum disorder, bipolar disorder, schizophrenia, and Parkinson's disease. In synthetic data, SVA usage frequently led to high false positive rates. This issue was not revealed by standard diagnostics nor resolved by strategies such as maximizing cross-dataset signal correlation. In real data, using SVA reduced cross-cohort reproducibility of differential expression signatures regardless of the number of surrogate variables included. These findings show that SVA risks inducing overfitting at realistic signal levels, advocating for a cautious, case-by-case application, with a preference for modeling known covariates.
ER -