regionalpcs improve discovery of DNA methylation associations with complex traits

Tiffany Eulalio; Min Woo Sun; Olivier Gevaert; Michael D Greicius; Thomas J Montine; Daniel Nachun; Stephen B Montgomery

doi:10.1038/s41467-024-55698-6

regionalpcs improve discovery of DNA methylation associations with complex traits

Nat Commun. 2025 Jan 3;16(1):368. doi: 10.1038/s41467-024-55698-6.

Authors

Tiffany Eulalio¹, Min Woo Sun², Olivier Gevaert^{2

3}, Michael D Greicius⁴, Thomas J Montine⁵, Daniel Nachun^#⁶, Stephen B Montgomery^#^{7

8}

Affiliations

¹ Department of Biomedical Data Science, Stanford University, Stanford, CA, USA. [email protected].
² Department of Biomedical Data Science, Stanford University, Stanford, CA, USA.
³ Stanford Center for Biomedical Informatics Research (BMIR), Department of Medicine, Stanford University, Stanford, CA, USA.
⁴ Department of Neurology & Neurological Sciences, Stanford University, Stanford, CA, USA.
⁵ Department of Pathology, Stanford University, Stanford, CA, USA.
⁶ Department of Pathology, Stanford University, Stanford, CA, USA. [email protected].
⁷ Department of Biomedical Data Science, Stanford University, Stanford, CA, USA. [email protected].
⁸ Department of Pathology, Stanford University, Stanford, CA, USA. [email protected].

^# Contributed equally.

Abstract

We have developed the regionalpcs method, an approach for summarizing gene-level methylation. regionalpcs addresses the challenge of deciphering complex epigenetic mechanisms in diseases like Alzheimer's disease. In contrast to averaging, regionalpcs uses principal components analysis to capture complex methylation patterns across gene regions. Our method demonstrates a 54% improvement in sensitivity over averaging in simulations, providing a robust framework for identifying subtle epigenetic variations. Applying regionalpcs to Alzheimer's disease brain methylation data, combined with cell type deconvolution, we uncover 838 differentially methylated genes associated with neuritic plaque burden-significantly outperforming conventional methods. Integrating methylation quantitative trait loci with genome-wide association studies identified 17 genes with potential causal roles in Alzheimer's disease risk, including MS4A4A and PICALM. Available in the Bioconductor package regionalpcs, our approach facilitates a deeper understanding of the epigenetic landscape in Alzheimer's disease and opens avenues for research into complex diseases.

MeSH terms

Alzheimer Disease* / genetics
Brain / metabolism
Brain / pathology
DNA Methylation* / genetics
Epigenesis, Genetic*
Genetic Predisposition to Disease
Genome-Wide Association Study*
Humans
Principal Component Analysis
Quantitative Trait Loci*

Abstract

MeSH terms

Grants and funding