Testing the Parsimony Test of Genome Duplications: A Counterexample
- 1Department of Zoology and Genetics, 2Center for Bioinformatics and Biological Statistics, and 3Department of Statistics, Iowa State University, Ames, Iowa 50011, USA
This extract was created in the absence of an abstract.
Whereas the role of genome duplication(s) in yeast and plants has been widely accepted, the hypothesis of genome duplication in early vertebrates (Ohno 1970) is still under controversy (Wolfe 2001). According to the current version, the 2R model, there were two rounds of polyploidization: one occurring before the divergence of jawless vertebrates and the other just after (Sidow 1996). Recently, doubt has been raised about the 2R model because the evidence was found to be weaker than previously thought (Wolfe 2001). For the proponents of the 2R model, this doubt may be explained as a combination of rapid gene deletion, sequence diversity, and chromosome rearrangement (Nadeau and Sankoff 1997; Wang and Gu 2000). For the opponents, however, the lack of strong evidence is sufficient to refute the 2R model, applying “Ockham's Razor” (Hughes et al. 2001; Makalowski 2001).
Alternatively, the model of small-scale tandem duplications (TDs) followed by translocations was invoked (Hughes et al. 2001). Moreover,Hughes et al. (2001) used the parsimony to test whether the TD hypothesis is “better” than the 2R hypothesis. The basic procedure is to infer the minimum number (G) of genetic events to explain the gene's current distribution on human chromosomes under each competing model. Here the genetic events include gene duplications (D), losses (L), and translocations (T), that is,GM = D + L + T, where the subscript M = 2R for the 2R model orTD for the TD model. Under this parsimony, the TD hypothesis is favored if GTD < G2R ; otherwise, the 2R model is favored. After examining 20 vertebrate gene families,Hughes …
