Genetic and neuro-epigenetic effects of divergent artificial selection for feather pecking behaviour in chickens

Abstract

Feather pecking (FP) is a repetitive behaviour in chickens, influenced by genetic, epigenetic, and environmental factors, similar to behaviours seen in human developmental disorders (e.g., hyperactivity, autism). This study examines genetic and neuro-epigenetic factors in the thalamus of chickens from lines selected for seven generations for high or low FP behaviour (HFP or LFP). We integrate data on Differentially Methylated Regions (DMRs), Single Nucleotide Polymorphisms (SNPs), and Copy Number Variations (CNVs) in this controlled artificial selection process. Significant differences in behaviour, immunology, and neurology have been reported in these lines. We identified 710 SNPs in these lines that indicate new potentially important genes for FP such as TMPRSS6 (implicated in autism), and SST and ARNT2 (somatostatin function). CNV were the omic level most affected during selection. The largest CNVs found were in RIC3 (gain in HFP) and SH3RF2 (gain in LFP) genes, linked to nicotinic acetylcholine receptor regulation and human oncogenesis, respectively. Our study also suggests that promoters and introns are hotspots for CpG depletion. The overlapping of the omic levels investigated here with data from a public FP Quantitative Trait Loci (QTL) database revealed novel candidate genes for understanding repetitive behaviours, such as RTKN2, associated with Alzheimer's disease in humans. This study suggests CNVs as a crucial initial step for genomic diversification, potentially more impactful than SNPs.

Keywords: Artificial selection; Chickens; Copy number variations; DNA methylation; Feather pecking; Genomics; Quantitative Traits Loci; Single nucleotide polymorphisms.

Fig. 1

Genetic separation of individuals in each lineage (9 HFP and 7 LFP), based on 76,414 SNPs remaining after filtering (sample call rate ≥ 20%; loci call rate ≥ 70%), employing a) PCA or b) Neighbour Joining distance

Fig. 2

a) Manhattan plot depicting the SNPs that crossed the adjusted P-values of p ≤ 0.05 (711 SNPs; red line) and p ≤ 0.0003 (46 SNPs; green line) after filtering by Benjamini-Hochberg false discovery rate (FDR); b) Representation of the allelic differences across the individuals on the HFP and LFP lineages obtained from 46 SNPs with p ≤ 0.0003

Fig. 3

Functional annotation of a) SNPs, b) CpG SNPs, and c) Novel CpGs found between the animals in the LFP and HFP lineages

Fig. 4

Schematic representation of the three loci showing differentially fixed alleles between HFP and LFP animals

Fig. 5

a) PCA showing genetic separation of individuals in each lineage (9 HFP and 7 LFP) based on CNVs; b) Manhattan plot showing CNV increases in each lineage across chromosomes; c) Functional annotation of CNVs found between LFP and HFP animals

Fig. 6

a) PCA showing methylomic separation of individuals in each lineage (9 HFP and 7 LFP) based on DMRs; b) Volcano plot showing fold changes of DMRs found between LFP and HFP animals; c) Functional annotation of DMRs found between LFP and HFP animals; d) Heat map showing methylation levels of the significant DMRs per individual

Fig. 7

a) Functional annotation of publicly available QTLs for feather pecking behaviour; b) Venn diagram depicting the total number of genetic and epigenetic difference between animals in estimated in this study, as well as the common among them and publicly available QTLs for feather pecking

Fig. 8

Graphical representation of overlaps of interest: a) the only region overlapping all the genomic levels investigated, which included a QTL for feather pecking, 33 CNVs, four SNPs and two adjacent DMRs (merged into one in the figure), located in the promoter of RTKN2; b) a DMR containing 21 CpGs that overlapped with one SNP, located in the promoter region of the novel gene ENSGALG00000032525

Fig. 9

Graphical representation of repeat masker analysis on the SNPs, CNVs and DMRs found between LFP and HFP animals