Surrogate variables and the risk of overfitting in neuropsychiatric gene expression analyses
Dung Hoang, Ryan Thompson, Anina Lund, Eli Kritzer, Ariela Buxbaum Grice, Divya John, Girish Nadkarni, Negar Golestani, Alexander William Charney, Noam Beckmann
This study shows that surrogate variable analysis (SVA) can lead to overfitting and reduced cross-cohort reproducibility in neuropsychiatric gene expression analyses.
bioRxiv · Oct 12, 2026 · PreprintBioinformatics