What are the consequences of using consensus trees in phylogenetic comparative analyses?

A phylogeny (evolutionary tree) is required for many different analytical techniques for data collected from several or many species.

Often, researchers go to the literature to find a phylogeny for their group, which is very often a consensus tree. However, from a Bayesian viewpoint, this is undesirable as consensus trees may have very little support from a Bayesian phylogenetic analysis. In fact, the consensus tree may not even be in the posterior tree set. Potentially, this could create problems for phylogenetic comparative analyses, as working with the wrong tree can provide misleading inferences.

This project will examine the above problem by:

• analysing previously published comparative studies using Bayesian methods, constructing consensus trees, determining their topology and comparing with trees with high posterior probability
• re-analysing published comparative studies using consensus trees and trees with high posterior probability to see if the different trees make a difference to the inference.

Supervisor: Dr Simon Blomberg