Exploring genome wide bisulfite sequencing for DNA methylation analysis in livestock: a technical assessment

Abstract

Recent advances made in "omics" technologies are contributing to a revolution in livestock selection and breeding practices. Epigenetic mechanisms, including DNA methylation are important determinants for the control of gene expression in mammals. DNA methylation research will help our understanding of how environmental factors contribute to phenotypic variation of complex production and health traits. High-throughput sequencing is a vital tool for the comprehensive analysis of DNA methylation, and bisulfite-based strategies coupled with DNA sequencing allows for quantitative, site-specific methylation analysis at the genome level or genome wide. Reduced representation bisulfite sequencing (RRBS) and more recently whole genome bisulfite sequencing (WGBS) have proven to be effective techniques for studying DNA methylation in both humans and mice. Here we report the development of RRBS and WGBS for use in sheep, the first application of this technology in livestock species. Important technical issues associated with these methodologies including fragment size selection and sequence depth are examined and discussed.

Keywords: DNA methylation; RRBS; WGBS; epigenetics; fragment size; quantification; sheep.

FIGURE 1

Base pair composition showing the first 10 bp from read one Illumina HiSeq 100 bp paired end sequencing indicating the expected MspI restriction site at the 5’ end of ~97% of fragments sequenced.

FIGURE 2

Bioanalyzer gel image of the three RRBS libraries made with different insert sizes. Ligated adapters cause the DNA fragments to migrate to a higher molecular weight (approximately 100 bp higher) than the insert sizes selected.

FIGURE 3

Comparison of coverage at genomic features (genes and promoter regions) from sequence data generated from libraries constructed from various fragment sizes (50–150, 150–250, and 250–350 bp). For inclusion, the genomic feature had to contain at least three CpG sites with ≥10x coverage. Promoter regions were defined as 2 kb upstream of the transcription start site.

FIGURE 4

Comparison of coverage at genomic features (genes and promoter regions) from sequence data generated from a reduced representation bisulfite sequencing (RRBS) library versus a whole genome bisulfite sequencing (WGBS) library. For inclusion, the genomic feature had to contain at least three CpG sites with ≥10x coverage. Promoter regions were defined as 2 kb upstream of the transcription start site.

FIGURE 5

CpG coverage generated from RRBS when smaller amounts of data are available for analysis. RRBS data were randomly sampled from the original fastq file to create smaller data sets, (A) Seqmonk screen shot illustrating CpG site coverage across selected genes on chromosome 1, bar height represents a count of CpG coverage, illustrating inclusions of increasing number of genes analyzed as the number of sequences included increases; (B) The number of CpGs covered in these sequentially smaller datasets was identified, in addition to the number of CpGs with at least 10x coverage.

FIGURE 6

CpGs with at least 10x coverage in data sets generated from RRBS versus WGBS.