Figure 1

Distributions of sum of branch lengths in a phylogenetic tree estimated from real data and synthetic data respectively. Sequences of eight Drosophila species were collected from real data ("Real"), data produced by a traditional simulator ("Traditional"), and data produced by the new simulator based on parameter sampling ("New"). The traditional simulator used the average substitution rates observed in the real data, while the new simulator used the empirical distribution of substitution rates in real data. The branch lengths were estimated by Paml [51].