Email updates

Keep up to date with the latest news and content from Biology Direct and BioMed Central.

Open Access Highly Accessed Research

Refuting phylogenetic relationships

James Bucknam1, Yan Boucher2 and Eric Bapteste1*

Author Affiliations

1 Canadian Institute for Advanced Research and Genome Atlantic, Department of Biochemistry and Molecular Biology, Dalhousie University, Halifax, Nova Scotia, B3H 4H7, Canada

2 Department of Chemistry and Molecular Biosciences, Macquarie University, North Ryde, NSW 2109, Australia

For all author emails, please log on.

Biology Direct 2006, 1:26  doi:10.1186/1745-6150-1-26

Published: 6 September 2006

Abstract

Background

Phylogenetic methods are philosophically grounded, and so can be philosophically biased in ways that limit explanatory power. This constitutes an important methodologic dimension not often taken into account. Here we address this dimension in the context of concatenation approaches to phylogeny.

Results

We discuss some of the limits of a methodology restricted to verificationism, the philosophy on which gene concatenation practices generally rely. As an alternative, we describe a software which identifies and focuses on impossible or refuted relationships, through a simple analysis of bootstrap bipartitions, followed by multivariate statistical analyses. We show how refuting phylogenetic relationships could in principle facilitate systematics. We also apply our method to the study of two complex phylogenies: the phylogeny of the archaea and the phylogeny of the core of genes shared by all life forms. While many groups are rejected, our results left open a possible proximity of N. equitans and the Methanopyrales, of the Archaea and the Cyanobacteria, and as well the possible grouping of the Methanobacteriales/Methanoccocales and Thermosplasmatales, of the Spirochaetes and the Actinobacteria and of the Proteobacteria and firmicutes.

Conclusion

It is sometimes easier (and preferable) to decide which species do not group together than which ones do. When possible topologies are limited, identifying local relationships that are rejected may be a useful alternative to classical concatenation approaches aiming to find a globally resolved tree on the basis of weak phylogenetic markers.

Reviewers

This article was reviewed by Mark Ragan, Eugene V Koonin and J Peter Gogarten.