Domestication affects sex-biased gene expression evolution in the duck

Abstract

Genes with sex-biased expression are thought to underlie sexually dimorphic phenotypes and are therefore subject to different selection pressures in males and females. Many authors have proposed that sexual conflict leads to the evolution of sex-biased expression, which allows males and females to reach separate phenotypic and fitness optima. The selection pressures associated with domestication may cause changes in population architectures and mating systems, which in turn can alter their direction and strength. We compared sex-biased expression and genetic signatures in wild and domestic ducks (Anas platyrhynchos), and observed changes of sexual selection and identified the genomic divergence affected by selection forces. The extent of sex-biased expression in both sexes is positively correlated with the level of both d _N /d _S and nucleotide diversity. This observed changing pattern may mainly be owing to relaxed genetic constraints. We also demonstrate a clear link between domestication and sex-biased evolutionary rate in a comparative framework. Decreased polymorphism and evolutionary rate in domesticated populations generally matched life-history phenotypes known to experience artificial selection. Taken together, our work suggests the important implications of domestication in sex-biased evolution and the roles of artificial selection and sexual selection for shaping the diversity and evolutionary rate of the genome.

Keywords: domestication; duck; sequence evolution; sex-biased expression; sexual selection.

Figure 1.

Sexual size dimorphism (SSD) is associated with domestication in ducks. dimorphism in ducks. SSD is calculated as the ratio between the average body weight of males compared to the total average weight (males and females). Z-tests for two independent samples were used so that individual weight data and sample size were both taken into consideration. Significant difference between SSD of three breeds is indicated (Z-test, *p < 0.05, **p < 0.01, ***p < 0.001).

Figure 2.

Heatmaps and hierarchical clustering of gene expression for (a) all samples; (b) gonad; and (c) liver. Shown is the average relative expression for all autosomal expressed genes from two tissues (L, liver; G, gonad) of M (male) and F (female) in three breeds (MD, PK and GY).

Figure 3.

Dynamic changes of sex-biased genes in three breeds. (a) Three categories of sex-biased expression changes among wild duck and domesticated duck. (b) A Venn diagram illustrating statistical results of sex-biased expression in three breeds. (c) The three categories of sex-biased changes and their numbers and proportions.

Figure 4.

Average ratio of nonsynonymous substitutions (d_N) to synonymous substitutions (d_S) for male-biased, female-biased and unbiased genes in three breeds. (a) Relationship between d_N/d_S and extent of sex-biased expression in gonads of three breeds, ‘high’, ‘medium’ and ‘low’ below the x-axis are shorthand for ‘high biased expression’, ‘medium biased expression’ and ‘low biased expression’, respectively. (b) Differences in d_N/d_S among breeds for three biased categories. Displayed significance scores are *p < 0.05, **p < 0.01 and ***p < 0.001.

Figure 5.

Amount of genetic diversity in 10 kb windows explained by the nucleotide diversity (π) for female-biased, unbiased, and male-biased genes expressed in gonads of three breeds. Genetic diversity also significantly varied across three breeds in the same bias categories (Wilcoxon Rank sum test, *p < 0.05, **p < 0.01, ***p < 0.001).

Figure 6.

Relationship between Tajima's D and extent of sex-biased expression in gonads of MD, PK and GY, ‘high’, ‘medium’ and ‘low’ below the x-axis are shorthand for ‘high biased expression’, ‘medium biased expression’ and ‘low biased expression’, respectively.