Identification of Genetic and Epigenetic Marks Involved in Population Structure

Jingyu Liu, Kent E. Hutchison, Nora I. Perrone‐Bizzozero, M.y. Morgan, Jing Sui, Vince D. CalhounPLoS 2010 · DOI

Abstract

Population structure is a well-recognized factor in genetic studies, yet its influence on epigenetics remains poorly understood. This study investigated the impact of population diversity on genome-wide single nucleotide polymorphisms (SNPs) and DNA methylation levels in 196 participants from five ethnic groups. Using principal and independent component analyses, researchers identified three population stratification factors (PSFs) in the genomic SNP dataset (6% of total variance) and one PSF in the genomic methylation dataset (0.2% of total variance). Notably, the methylation PSF significantly correlated with the primary SNP PSF ($r = 0.72, p < 10^{-23}$). The SNP PSF was primarily composed of genes associated with pigmentation (SLC45A2, HERC2, CTNNA2), while the methylation PSF involved 48 sites across 44 genes related to basic molecular functions. These findings suggest that genetic and epigenetic population structures are linked through complex gene interactions and shared biological processes, potentially mediated by the ID2 repressor.

Introduction

In Genome-Wide Association Studies (GWAS), population structure—the systematic difference in allele frequencies between sub-populations due to different ancestry—is a critical confounding variable. Failure to account for these differences can lead to spurious associations. While robust methods exist to control for population structure in genetics, the extent and nature of "epigenetic population structure" are less clear. DNA methylation, a key epigenetic mark, is influenced by both genetic sequence and environmental factors. This study aims to characterize the relationship between genetic ancestry and DNA methylation patterns across diverse ethnic groups and identify the biological mechanisms bridging the two.

Methods

The study analyzed saliva-derived DNA from 196 individuals representing five self-reported ethnic groups: Caucasian ($n=90$), Latino ($n=52$), Mixed ($n=37$), Native American ($n=12$), African American ($n=4$), and Asian ($n=1$).

Genetic variation was assessed using Illumina Human 1M-Duo SNP arrays, while epigenetic variation was measured using Illumina Infinium 27K methylation arrays. To identify population stratification factors (PSFs), the researchers applied Principal Component Analysis (PCA) to the SNP data and Independent Component Analysis (ICA) to the methylation data. ICA was chosen for the epigenetic dataset due to its superior ability to isolate independent biological signals from complex, noisy data. Correlation analyses and pathway enrichment tools (DAVID and MetaCore) were subsequently used to map the relationships between genetic and epigenetic markers.

Results

The Genetic-Epigenetic Correlation

The analysis revealed a stark difference in the variance explained by population structure: 6% for genetic data versus only 0.2% for epigenetic data. This indicates that while ancestry is a primary driver of genetic variation, DNA methylation is more heavily influenced by non-ancestral factors such as age, lifestyle, and environment.

Despite this "variance gap," the primary genetic PSF and the primary epigenetic PSF were highly correlated ($r = 0.72, p < 10^{-23}$). This strong link suggests that a subset of epigenetic marks is tightly coupled with genetic ancestry.

Key Markers and Biological Pathways

The top contributors to the genetic PSF were eight SNPs within three genes—SLC45A2, HERC2, and CTNNA2—all of which are well-known for their roles in determining skin, hair, and eye color. In contrast, the methylation PSF comprised 48 sites in 44 genes involved in fundamental molecular processes, including transcription regulation, DNA binding, and cytokine activity.

Eight specific methylation sites were found to correlate directly with the minor alleles of SNPs in the genetic PSF. For instance, the SNP rs16891982 in SLC45A2 was strongly associated with the methylation status of the PM20D1 gene. Functional analysis indicated that the genes in both PSFs share common biological pathways, such as multicellular organism reproduction, cell-cell signaling, and cytoskeleton organization.

The ID2 Regulatory Hub

Network analysis identified ID2 (Inhibitor of DNA Binding 2) as a central regulatory node. Most genes identified in the PSFs interact closely with ID2, which encodes a helix-loop-helix protein that inhibits DNA binding. This suggests that genetic variations in ancestry-linked pigmentation genes may modulate ID2 function, which then acts as a repressor to influence the methylation levels of a broad network of genes.

Discussion

This study demonstrates a significant correlation between genetic and epigenetic population stratification. The findings suggest that the interrelationship is not random but mediated through specific biological processes and regulatory hubs like ID2.

However, several limitations must be noted. The sample size for certain ethnic groups (Asian and African American) was very small, limiting generalizability. Additionally, the use of saliva—a heterogeneous tissue—means results may vary in other tissues due to the high tissue-specificity of DNA methylation. Finally, while the study establishes a strong correlation, further functional studies are required to prove a causal link between the identified SNPs and the observed methylation patterns. Future research should focus on larger, more diverse cohorts and longitudinal data to better understand how environmental factors interact with these ancestry-linked epigenetic marks.

Summary updated

Abstract

Population structure is well known as a prevalent and important factor in genetic studies, but its relevance in epigenetics is unclear. Very little is known about the affected epigenetic markers and their connections with genetics. In this study we assessed the impact of population diversity on genome wide single nucleotide polymorphisms (SNPs) and DNA methylation levels in 196 participants from five ethnic groups, using principle and independent component analyses. Three population stratification factors (PSFs) were identified in the genomic SNP dataset, accounting for a relatively large portion of total variance (6%). In contrast, only one PSF was identified in genomic methylation dataset accounting for 0.2% of total variance. This methylation PSF, however, was significantly correlated with the largest SNP PSF (r = 0.72, p<1E-23). We then investigated the top contributing markers in these two linked PSFs. The SNP PSF predominantly consists of 8 SNPs from three genes, SLC45A2, HERC2 and CTNNA2, known to encode skin/hair/eye color. The methylation PSF includes 48 methylated sites in 44 genes coding for basic molecular functions, including transcription regulation, DNA binding, cytokine, and transferase activity. Among them, 8 sites are either hypo- or hyper-methylated correlating to minor alleles of SNPs in the SNP PSF. We found that the genes in SNP and methylation PSFs share common biological processes including sexual/multicellular organism reproduction, cell-cell signaling and cytoskeleton organization. We further investigated the transcription regulatory network operating at these genes and identified that most of genes closely interact with ID2, which encodes for a helix-loop-helix inhibitor of DNA binding. Overall, our results show a significant correlation between genetic and epigenetic population stratification, and suggest that the interrelationship between genetic and epigenetic population structure is mediated via complex multiple gene interactions in shared biological processes, through possibly, SNP-dependent modulation and ID2 repressor function.
BibTeX
@inproceedings{liu2010identification,
  author = {Jingyu Liu and Kent E. Hutchison and Nora I. Perrone‐Bizzozero and M.Y. Morgan and Jing Sui and Vince D. Calhoun},
  title = {{Identification of Genetic and Epigenetic Marks Involved in Population Structure}},
  booktitle = {PLoS},
  year = {2010},
  doi = {10.1371/journal.pone.0013209},
}
Figure