DNA methylation analysis to differentiate reference, breed, and parent-of-origin effects in the bovine pangenome era

Callum MacPhillamy, Tong Chen, Stefan Hiendleder, J. L. Williams, Hamid Alinejad‐Rokny, Wai Yee LowGigaScience 2024 · DOI

TL;DR

This study highlights a critical bias in DNA methylation analysis: the choice of reference genome. By using a bovine pangenome approach with Angus and Brahman cattle, researchers show that reference genome choice can significantly impact results, leading to inaccurate conclusions about breed and parent-of-origin effects. This work emphasizes the need for more comprehensive reference genomes to accurately study genetic and epigenetic variations.

What Problem Does This Paper Address?

DNA methylation is a crucial epigenetic mechanism that regulates gene expression. Most studies analyzing DNA methylation rely on a single reference genome. However, this approach can introduce significant bias, especially when studying genetically diverse populations or breeds. Genetic variations like single nucleotide polymorphisms (SNPs) and structural variants (SVs) can alter CpG sites (regions where methylation commonly occurs), leading to differences in methylation patterns that are misinterpreted as biological differences rather than artifacts of the reference genome used. This study aims to quantify this bias and demonstrate how using a pangenome approach can provide a more accurate understanding of DNA methylation differences related to breed and parent-of-origin effects in cattle.

Methods

The researchers employed whole-genome bisulfite sequencing (WGBS) on fetal liver samples from Angus (Bos taurus taurus) and Brahman (Bos taurus indicus) cattle, including reciprocal crosses. Key methods include:

  • WGBS Data Mapping: Reads were mapped to both the Angus and Brahman reference genomes separately.
  • Pangenome Approach: The study utilized reference genomes from the Bovine Pangenome Consortium, representing a more comprehensive view of genetic variation than a single reference.
  • Identification of Shared and Breed-Specific CpGs: They identified CpG sites common to both breeds and those unique to one, accounting for SNPs and SVs.
  • Quantification Bias Assessment: Statistical tests (Kolmogorov-Smirnov, Wilcoxon rank-sum) were used to measure differences in global methylation when using different reference genomes.
  • Differentially Methylated Region (DMR) Analysis: The methylKit package was used to identify regions with significant methylation differences between groups.
  • Gene Expression Analysis: RNA-sequencing data was used to identify differentially expressed genes (DEGs).
  • Overlap Analysis: The study examined the overlap between DMRs and DEGs, as well as imprinted genes.

Datasets

The study utilized whole-genome bisulfite sequencing (WGBS) data from 24 fetal liver samples. These samples represented four genetic groups:

  • BTBT: Purebred Angus (Angus sire, Angus dam)
  • BTBI: Angus sire, Brahman dam
  • BIBT: Brahman sire, Angus dam
  • BIBI: Purebred Brahman (Brahman sire, Brahman dam)

The study also incorporated RNA-sequencing data from the same samples. The reference genomes used were the Angus and Brahman reference genomes from the Bovine Pangenome Consortium.

Key Insights and Ideas

  • Reference Genome is Not Neutral: The choice of reference genome is not a neutral decision and can introduce significant, quantifiable bias into DNA methylation analyses.
  • Genetic Variation Drives Epigenetic Differences: SNPs and SVs are not just genetic differences; they directly impact CpG sites, creating "genetics of epigenetics" that can be misinterpreted without proper context.
  • Pangenomes are Essential for Divergent Populations: For species with significant genetic divergence between breeds (like Angus and Brahman), a single reference genome is insufficient. A pangenome approach is crucial for accurate interpretation.
  • DMRs May Act on Enhancers: Differentially methylated regions (DMRs) show a tendency to overlap with enhancer regions more than gene promoters, suggesting a key role for enhancers in mediating epigenetic effects on gene expression.
  • Parent-of-Origin Effects are Complex: The study provides evidence for parent-of-origin effects on methylation, but disentangling these from breed effects requires careful consideration of the reference genome.

Key Results

  • Approximately 75% of CpG sites were shared between Angus and Brahman, with ~5% being breed-specific and ~20% unresolved.
  • Using an incorrect reference genome introduced up to ~2% quantification bias in global methylation.
  • SNPs impacted CpGs 13 times more than other autosomal sites.
  • Structural variants (SVs) contained 1.18 times more CpGs than non-SVs.
  • There was a poor overlap between DMRs and DEGs, suggesting DMRs might influence gene expression indirectly, possibly through enhancers.
  • DMRs overlapped with imprinted genes, including DGAT1, which is important for fat metabolism and weight gain, showing breed-specific and sire-of-origin effects.
  • Global CpG methylation levels were generally consistent with previous studies in mammals (47-62%).
  • The study found that breed-specific CpGs were significantly more likely to have lower methylation than shared CpGs.

Limitations and Future Work

  • Reference Genome Limitations: Even with pangenome resources, ~20% of CpGs were unresolved, indicating that current pangenomes may not capture all genetic variation. The study acknowledges that neither reference genome likely captures all variation present within the two breeds.
  • Short-Read Sequencing: The reliance on short-read sequencing limits the ability to fully resolve complex structural variants and definitively assign parent-of-origin effects without long-read sequencing.
  • Future Work: The authors suggest that personalized reference genomes or pangenome-based approaches are essential for accurate analyses, especially in livestock populations with high genetic divergence. They also highlight the potential of long-read sequencing to simplify DNA methylation analyses and improve accuracy. Further investigation into the role of enhancer methylation in embryonic and fetal development is also recommended.

Summary updated

Abstract

This work demonstrates the need to consider reference genome effects to explore genetic and epigenetic differences accurately and identify DMRs involved in controlling certain genes.
BibTeX
@inproceedings{macphillamy2024dna,
  author = {Callum MacPhillamy and Tong Chen and Stefan Hiendleder and J. L. Williams and Hamid Alinejad‐Rokny and Wai Yee Low},
  title = {{DNA methylation analysis to differentiate reference, breed, and parent-of-origin effects in the bovine pangenome era}},
  booktitle = {GigaScience},
  year = {2024},
  doi = {10.1093/gigascience/giae061},
}
Figure