Tissue Specificity and Dynamics of Sex-Biased Gene Expression in a Common Frog Population with Differentiated, Yet Homomorphic, Sex Chromosomes

Abstract

Sex-biased genes are central to the study of sexual selection, sexual antagonism, and sex chromosome evolution. We describe a comprehensive de novo assembled transcriptome in the common frog Rana temporaria based on five developmental stages and three adult tissues from both sexes, obtained from a population with karyotypically homomorphic but genetically differentiated sex chromosomes. This allows the study of sex-biased gene expression throughout development, and its effect on the rate of gene evolution while accounting for pleiotropic expression, which is known to negatively correlate with the evolutionary rate. Overall, sex-biased genes had little overlap among developmental stages and adult tissues. Late developmental stages and gonad tissues had the highest numbers of stage- or tissue-specific genes. We find that pleiotropic gene expression is a better predictor than sex bias for the evolutionary rate of genes, though it often interacts with sex bias. Although genetically differentiated, the sex chromosomes were not enriched in sex-biased genes, possibly due to a very recent arrest of XY recombination. These results extend our understanding of the developmental dynamics, tissue specificity, and genomic localization of sex-biased genes.

Keywords: adult tissues; development; gene expression; pleiotropy; rate of evolution; sex bias; sex chromosomes; tissue specificity.

Figure 1

Summary of tissue specificity of the most expressed genes at different expression thresholds (top 10%, 30%, 70%, and 90%) of all eight studied tissues/stages.

Figure 2

Summary of sex-biased gene expression Log₂FC (fold change) in all eight tissues/stages. The dashed line defines the Log₂FC threshold beyond which a gene was considered sex biased, with >1 indicating male bias and <−1 for female bias.

Figure 3

Breakdown of reads per kilobase million (RPKM) by sex bias and sample sex, for the five tissues/stages with >100 sex-biased genes. A: G43, B: G46, C: brain, D: gonad, and E: liver. p values of Wilcoxon rank sum tests are summarized above and below the box plots (NS = non-significant, *** = p < 0.001, ** = p < 0.005, * = p < 0.05).

Figure 4

Average female- and male-biased gene expression in bins of different sex-bias fold change thresholds for (A1, B1) G43, (A2, B2) G46, (A3, B3) brain, (A4, B4) gonad, and (A5, B5) liver. The significance level of Wilcoxon rank sum tests between successive bins of male or female-biased genes is indicated above and below the boxplots (NS = non-significant, *** = p < 0.001, ** = p < 0.005, * = p < 0.05, . = p < 0.1). The statistical tests amongst negative numbers are not biologically meaningful, as all negative log₂FC values represent raw RPKM between 0 and 1.

Figure 4

Average female- and male-biased gene expression in bins of different sex-bias fold change thresholds for (A1, B1) G43, (A2, B2) G46, (A3, B3) brain, (A4, B4) gonad, and (A5, B5) liver. The significance level of Wilcoxon rank sum tests between successive bins of male or female-biased genes is indicated above and below the boxplots (NS = non-significant, *** = p < 0.001, ** = p < 0.005, * = p < 0.05, . = p < 0.1). The statistical tests amongst negative numbers are not biologically meaningful, as all negative log₂FC values represent raw RPKM between 0 and 1.

Figure 5

The ratios of non-synonymous to synonymous substitutions dN/dS in the tissues with >100 sex-biased genes. Significant differences based on Wilcoxon rank sum tests are denoted (NA = not applicable (due to low sample size), NS = non-significant, *** = p < 0.001, ** = p < 0.005, * = p < 0.05).

Figure 6

Correlation between sex-bias and tissue specificity in the five tissues with >100 sex biased genes (A) G43, (B) G46, (C) brain, (D) gonad, and (E) liver.

Figure 7

Tissue specificity in both sex-biased and unbiased genes in the five tissues with >100 sex-biased genes. Significant differences based on Wilcoxon rank sum tests are denoted (NS = non-significant, *** = p < 0.001, ** = p < 0.005, * = p < 0.05).