Molecular evolution and the incorporation of site-to-site rate variation in distance tree construction methods

The construction of evolutionary trees based on sequence data is not self-e vident Apart from the plethora of methods and software tools to choose from if one wants to infer phylogenetic tree topologies, one has also to be cau tious about the sequence data themselves. In this paper, we discuss how sys tematic errors can be introduced by one of the phenomena that often charact erize sequence data, i.e. differences in substitution rates among the diffe rent sites of the molecule. Regarding painwise distance methods, these syst ematic errors can often be avoided if an appropriate substitution model is applied to the construction of phylogenetic trees. This is demonstrated for a phylogeny based on animal small subunit ribosomal RNA sequences.