Fig. 4.

Rate differences between singleton and duplicable genes are explained by CDS length and expression level differences. Analysis of 1,473 singletons (blue) and 70 duplicable genes (orange). Boxplot of (a) log₁₀ of expression levels + 1 (log2-transformed RPKM gene expression data) and (c) log₁₀ of CDS lengths of duplicable and singleton genes. P values for the test of difference between the means is shown above the boxplots (P values 0.237 and 0.001, respectively, MWU). LOWESS regression fitted for $d_N/d_S$ on the log₁₀ of (b) expression levels + 1 and (d) CDS length. The regression lines for singletons are shown in black and, for duplicates in orange. (e) $d_N/d_S$ residual space is graphed by plotting the residuals extracted from the two LOWESS regressions. The median residual value of the duplicable set is indicated by a large white point and the singleton set by a large black point. The dotted lines from each axis indicate the medians and 95 limits for the respective sets of residuals.