Figure 2
Human metabolic genes are under greater selective constraint than other orthologous genes. A) The distribution ω for metabolic genes (red, nm = 1,190), non-metabolic genes (black, no = 12,738), and the total set of orthologs (white, nt = nm+no= 13,922). The x-axis gives ω: the ratio of non-synonymous to synonymous substitutions per site (i.e., our proxy for selective constraint, see main text). The y-axis is the number of genes with a given ω value for each gene set. We can reject the hypothesis that the median of metabolic genes is not significantly smaller than the median of non-metabolic genes (P = 0.035; Mann-Whitney U-test). B) The white portion of the circle graph shows the relative proportions of genes for which we cannot identify orthologs for the metabolic and non-metabolic genes (red and black, respectively). These proportions are significantly different (χ2 = 11.98, P < 0.001). C) After taking samples of size n = 1,190 and n = 12,738 and calculating the difference in mean ω, 100,000 times, we found no case in which the absolute difference in mean values was as high as the observed difference in mean ω for metabolic and non-metabolic genes (P < 10-5).