Extended Data Fig. 7: Relationships between X-Y gemetolog expression/co-expression and functional divergence. | Nature Ecology & Evolution

Extended Data Fig. 7: Relationships between X-Y gemetolog expression/co-expression and functional divergence.

From: Evolutionary divergence between homologous X–Y chromosome genes shapes sex-biased biology

Extended Data Fig. 7

a. X–Y co-expression values in the current study versus those estimated by Godfrey and colleagues18. We recomputed X–Y co-expression values to be comparable to our estimates of functional divergence (aCFDXXF–XYM and sCFDXXF–XYM). Differences between X–Y co-expression values across studies are the result of methodological differences, as our study: i) used a newer version of the GTEx data (v8 versus v7); ii) removed age effects when calculating adjusted expression levels (Methods); and iii) normalized co-expression values by expression level using spatial quantile normalization (Methods). Regression lines are provided for each tissue (see legend) and across all data points (black line, confidence interval is shaded). These values are highly similar (ρ = 0.617, p < 2.2e-16; linear model: slope = 0.418; p < 2.2e-16) across pair–tissue combinations. Spearman’s rank order correlation and the corresponding p-value are provided. b. aCFDXXF–XYM versus X–Y member co-expression. Regression lines are provided for each tissue (see legend) and across all data points (black line, confidence interval is shaded). Absolute within-pair divergence (aCFDXM–YM) necessarily exhibits a strong negative correlation with the level of co-expression between the X and Y members themselves (within each pair) (ρ = -0.768, p < 2.2e-16). Spearman’s rank order correlation and the corresponding p-value are provided. c. sCFDXXF–XYM versus X–Y member co-expression. Regression lines are provided for each tissue (see legend) and across all data points (black line, confidence interval is shaded). Note that relative to aCFDXXF–XYM (Extended Data Fig. 7b), this relationship is relatively muted for signed divergence (sCFDXM–YM) (ρ = -0.172, p < 0.001). Spearman’s rank order correlation and the corresponding p-value are provided. d. aCFDXM–YM versus the median Y/X expression ratio. Regression lines are provided for each tissue (see legend) and across all data points (black line, confidence interval is shaded). Spearman’s rank order correlation and the corresponding p-value are provided. e. sCFDXM–YM versus the median Y/X expression ratio. Regression lines are provided for each tissue (see legend) and across all data points (black line, confidence interval is shaded). Spearman’s rank order correlation and the corresponding p-value are provided. f. Differences in X–Y expression variance versus differences in X–Y mean expression. Regression lines are provided for each tissue (see legend) and across all data points (black line, confidence interval is shaded). Spearman’s rank order correlation and the corresponding p-value are provided. g. Extended Data Fig. 7d excluding TMSB4X/Y. Regression lines are provided for each tissue (see legend) and across all data points (black line, confidence interval is shaded). Spearman’s rank order correlation and the corresponding p-value are provided. h. Extended Data Fig. 7e excluding TMSB4X/Y. Regression lines are provided for each tissue (see legend) and across all data points (black line, confidence interval is shaded). Spearman’s rank order correlation and the corresponding p-value are provided. i. Extended Data Fig. 7f excluding TMSB4X/Y. Regression lines are provided for each tissue (see legend) and across all data points (black line, confidence interval is shaded). j. Y/X expression ratios (log2) across all pair–tissue combinations. Purple = Y > X expression. Red = X > Y expression.

Back to article page