Stumped by Trees? A Generalized Null Model for Patterns of Organismal Diversity

Evolutionary biologists increasingly have become interested in the factors determining the structure of phylogenetic trees. For example, highly asymmetric trees seem to suggest that the probability of extinction and/or speciation differs among lineages. Before looking for the cause of such differences, one must establish that the structure of the trees differs significantly from results of an appropriate null model. A commonly invoked null model assumes that speciation is equally probable along all branches of a diversifying tree and predicts that highly asymmetric trees are not unexpected. However, this model of diversification assumes more than equality of rates among lineages. We demonstrate theoretically and via simulations that relaxation of the hidden assumptions that speciation is essentially instantaneous and occurs independently in separate lineages leads to qualitatively different expectations concerning the expected phylogenetic topology. In particular, highly asymmetric trees are considerably less likely to arise by chance when the speciation process takes up a significant fraction of the expected time between initiation of speciation events or when lineages tend to speciate simultaneously. As with most null models in ecology and evolution, the assumptions encoded in the model play an important role in determining the null expectation.