Pattern of divergence of amino acid sequences encoded by paralogous genes in human and pufferfish

We used phylogenetic analyses of protein families containing two or more pairs of orthologues in the genomes of human and pufferfish (Takifugu rubripes) to test the hypothesis that these sequences show a strong signal of polyploidization events hypothesized to have occurred early in vertebrate history. In order to test for evidence of two distinct rounds of polyploidization (the 2R hypothesis), we compared the pattern of amino acid sequence divergence of proteins encoded by genes duplicated just prior to the most recent common ancestor of human and pufferfish with that of proteins encoded genes duplicated earlier. These sequence divergences were statistically indistinguishable, contrary to the prediction of the 2R hypothesis. The variance of amino acid sequence divergences between paralogues was significantly greater than expected from that of orthologues in the same families. Estimation of gene duplication times assuming a molecular clock provided earlier estimates than expected, suggesting that it may not be appropriate to time the duplication of paralogues using rate estimates derived from orthologous comparisons. Overall, the results indicate that amino acid sequences do not provide a strong signal supporting the hypothesis that gene duplications early in vertebrate history occurred by polyploidization. On the other hand, the data are easily explained under an alternative model that gene duplications occurred at different times in different vertebrate gene families.