A chicken DNA methylation clock for the prediction of broiler health

Abstract

The domestic chicken (Gallus gallus domesticus) is the globally most important source of commercially produced meat. While genetic approaches have played an important role in the development of chicken stocks, little is known about chicken epigenetics. We have systematically analyzed the chicken DNA methylation machinery and DNA methylation landscape. While overall DNA methylation distribution was similar to mammals, sperm DNA appeared hypomethylated, which correlates with the absence of the DNMT3L cofactor in the chicken genome. Additional analysis revealed the presence of low-methylated regions, which are conserved gene regulatory elements that show tissue-specific methylation patterns. We also used whole-genome bisulfite sequencing to generate 56 single-base resolution methylomes from various tissues and developmental time points to establish an LMR-based DNA methylation clock for broiler chicken age prediction. This clock was used to demonstrate epigenetic age acceleration in animals with experimentally induced inflammation. Our study provides detailed insights into the chicken methylome and suggests a novel application of the DNA methylation clock as a marker for livestock health.

Fig. 1. A conserved DNA methylation system in chicken.

a Conservation of canonical DNMT enzymes in chicken. Conserved domains of animal DNMTs are shown in different colors. All DNMTs: catalytic domain (red). DNMT1: DMAP1 binding domain (dark blue), replication foci targeting domain (orange), CXXC domain (turquoise), BAH domain (green). DNMT3: PWWP domain (pink), ADD domain (purple). DNMT3L is a catalytically inactive DNMT3 variant that lacks the N-terminal part of the regulatory domain (including the PWWP domain) and the C-terminal part of the catalytic domain. b Evolutionary conservation of DNMT3L in vertebrates. The phylogenetic tree is based on 421 automatically annotated genomes from NCBI/refseq. c Interspecific comparison of somatic methylation landscapes. Violin plots show methylation ratios of 2 kb sliding windows covering the entire genome. The uneven plot shape for the elephant shark (Callorhinchus milii) is related to the relatively low sequencing coverage. The DNMT homologs of individual organisms are indicated under the respective species names.

Fig. 2. Methylation analysis of genomic features.

a Violin plots showing methylation levels of different genomic features in three different tissues. All differences between tissues are statistically significant (P = 2.2 × 10⁻¹⁶, Wilcoxon rank sum test), except the difference between lung and breast muscle at 3’-UTRs. b Violin plots showing methylation levels of different repeat classes in three different tissues. c Representative gene body methylation pattern. The methylation track shows DNA methylation levels (light blue) and sequencing coverage levels (red) for all CpGs of the chicken DNMT3B locus (chromosome 20) in lung. Sequencing coverages were cut off at >10. d Violin plots showing methylation ratios of 2 kb sliding windows. Representative somatic datasets are from chicken lung and mouse intestine.

Fig. 3. Identification of low-methylated regions (LMRs) in the chicken methylome.

a Size distribution of chicken LMRs. b Principal component analysis of LMR methylation patterns from different tissues, strains, and age groups. c Age-related differential methylation of LMRs in the jejunum of broiler chickens. Representative methylation tracks for the PDIA4 gene in d14 (blue) and d28 (crimson) samples. The position of LMRs is indicated by purple horizontal bars. Note the pronounced differential methylation between d14 and d28 for the rightmost LMR.

Fig. 4. Establishment and validation of a multi-tissue DNA methylation clock for broiler chicken.

a PCA analysis based on the DNA methylation patterns of 67,651 LMRs in the training set (N = 36 samples). b PCA based on LMR methylation patterns after filtering and normalization. c Example of a highly weighted clock LMR from the Igfbp3 gene body (#5 in Tab. S4), showing progressive age-related methylation. Coordinates refer to chromosome 2 of chicken genome assembly (version 5.0). Green lines indicate methylation levels from 0 to 1. d Enrichment of clock LMRs and clock CpGs in specific genomic features. P promoters, G gene bodies, I intergenic regions. e Validation of LMR-based and CpG-based age prediction in breast muscle tissue from 6 independent animals belonging to 2 age groups (14 days and 28 days). Error bars indicate RMSEs.

Fig. 5. Intestinal inflammation is associated with age acceleration in broiler chickens.

a Schematic outline of the inflammation trial. After hatching (day 0), animals were periodically injected with CpG (or control GpC) oligonucleotide, as shown (orange tringles). At several timepoints (black circles), animals were euthanized and tissue was collected for analysis. b Inflammatory kinase activation in jejunum samples with experimentally induced inflammation. The number of observed differentially phosphorylated proteins in the top 10 immune related GO terms were counted for CpG injected broiler chickens at each day. At day 35 post-hatch, a substantial reduction in protein phosphorylation changes related to immunity was observed. c Age acceleration in jejunum samples with experimentally induced inflammation. The specific age acceleration in the d13-16 age group matches the immunological phenotype. Asterisks indicate statistically significant differences in relation to day 35 (P < 0.05, t-test).