Rapid subfunctionalization accompanied by prolonged and substantial neofunctionalization in duplicate gene evolution

Abstract

Gene duplication is the primary source of new genes. Duplicate genes that are stably preserved in genomes usually have divergent functions. The general rules governing the functional divergence, however, are not well understood and are controversial. The neofunctionalization (NF) hypothesis asserts that after duplication one daughter gene retains the ancestral function while the other acquires new functions. In contrast, the subfunctionalization (SF) hypothesis argues that duplicate genes experience degenerate mutations that reduce their joint levels and patterns of activity to that of the single ancestral gene. We here show that neither NF nor SF alone adequately explains the genome-wide patterns of yeast protein interaction and human gene expression for duplicate genes. Instead, our analysis reveals rapid SF, accompanied by prolonged and substantial NF in a large proportion of duplicate genes, suggesting a new model termed subneofunctionalization (SNF). Our results demonstrate that enormous numbers of new functions have originated via gene duplication.

Figure 1.—

Evolutionary models of functional divergence between duplicate genes. The three neofunctionalization (NF) models differ in the number of ancestral functions retained by the gene that acquires novel functions. The newly proposed subneofunctionalization (SNF) model is a combination of NF and subfunctionalization (SF). Duplicate genes are depicted by open circles and different gene functions are shown by solid squares. Dotted lines link genes with their functions. In this article, we analyze functions of duplicate genes by their protein interaction partners and expression sites.

Figure 2.—

Number of protein interaction partners of yeast duplicate and singleton genes. (a) Mean number of partners (t) for duplicate pairs with different d_S. The error bar shows one standard error of the mean. The dashed line shows the average number of partners per singleton (A) and the dotted line shows the average number per random pair of singletons (T). There are 17, 15, 70, and 229 duplicate pairs in the four bins, respectively. (b) Frequency distributions of the number of interaction partners for singletons (open bars), singleton pairs (shaded bars), and duplicates (solid bars). (c) Frequency distributions of the number of shared partners between singleton pairs (open bars), duplicates with d_S < 20 (shaded bars), and duplicates with d_S > 20 (solid bars). (d) Difference in the number of partners between duplicates. The mean max(a₁, a₂), min(a₁, a₂), and δ = |a₁ − a₂| for duplicates are shown in open, shaded, and solid bars, respectively, with their corresponding mean values from random singleton pairs shown by the dotted, dashed, and solid lines, respectively. For any given pair of duplicates, a₁ and a₂ are the numbers of partners that they each have.

Figure 3.—

Number of expression sites of human duplicate and singleton genes. (a) Mean number of expression sites (t, squares) and mean number of shared expression sites (s, circles) for duplicate pairs with different d_S. The 1230 duplicate gene pairs are ranked by their d_S. Each square (or circle) represents the median d_S and mean t (or mean s) for 100 gene pairs, with the exception of the last square (or circle), which is derived from 630 gene pairs. The error bar shows one standard error of the mean. The dashed line shows the average number of expression sites per singleton (A) and the dotted line shows the average number per random pair of singletons (T). The solid line shows the average number of shared expression sites per random pair of singletons. (b) Difference in the number of expression sites between duplicates. The mean δ = |a₁ − a₂| and min(a₁, a₂) for duplicates are shown in squares and circles, respectively, with their corresponding mean values from singleton pairs shown by the dotted and dashed lines, respectively. For any given pair of duplicates, a₁ and a₂ are the numbers of expression sites that they each have.