Skip to content

CONFIDENCE LIMITS ON PHYLOGENIES: AN APPROACH USING THE BOOTSTRAP

Why this mattered

Felsenstein’s 1985 paper changed phylogenetics by making uncertainty in evolutionary trees a routine object of analysis rather than an informal caveat. Before this, a published tree often appeared as a single best estimate, with support judged through character counts, congruence, or expert assessment. By adapting Efron’s bootstrap to phylogenetic data, Felsenstein gave systematists a practical way to ask which inferred clades were stable under perturbations of the observed characters. The key conceptual move was to resample characters, not taxa: the species remained fixed, while sites or traits were treated as repeated observations bearing information about the underlying history. This reframed a phylogeny from a static diagram into a statistical estimate with measurable sampling error.

The immediate practical consequence was that confidence-like support values could be attached to branches using existing tree-building programs, simply by reweighting characters according to how often they appeared in bootstrap replicates. That made large-scale comparative phylogenetics more auditable: researchers could distinguish well-supported monophyletic groups from fragile groupings produced by limited or conflicting data. The paper also helped standardize majority-rule consensus trees as a way to summarize many inferred phylogenies, turning computational repetition into interpretable evidence. Its influence was especially large because it was methodologically general: it could be used with parsimony and later with distance, likelihood, and other tree-estimation frameworks.

Its broader importance lies in making statistical resampling part of the grammar of evolutionary inference. As molecular sequence data expanded through the late twentieth century, bootstrap support became one of the most recognizable measures in phylogenetic papers, appearing beside branches in trees across systematics, genomics, epidemiology, and comparative biology. Later developments refined or challenged aspects of bootstrap interpretation, including model-based phylogenetics, Bayesian posterior probabilities, approximate likelihood-ratio tests, and coalescent methods for gene-tree/species-tree discordance. But those later breakthroughs inherited the expectation that a phylogeny should not merely be proposed; its internal claims should be accompanied by explicit, reproducible measures of support.

Abstract

The recently-developed statistical method known as the "bootstrap" can be used to place confidence intervals on phylogenies. It involves resampling points from one's own data, with replacement, to create a series of bootstrap samples of the same size as the original data. Each of these is analyzed, and the variation among the resulting estimates taken to indicate the size of the error involved in making estimates from the original data. In the case of phylogenies, it is argued that the proper method of resampling is to keep all of the original species while sampling characters with replacement, under the assumption that the characters have been independently drawn by the systematist and have evolved independently. Majority-rule consensus trees can be used to construct a phylogeny showing all of the inferred monophyletic groups that occurred in a majority of the bootstrap samples. If a group shows up 95% of the time or more, the evidence for it is taken to be statistically significant. Existing computer programs can be used to analyze different bootstrap samples by using weights on the characters, the weight of a character being how many times it was drawn in bootstrap sampling. When all characters are perfectly compatible, as envisioned by Hennig, bootstrap sampling becomes unnecessary; the bootstrap method would show significant evidence for a group if it is defined by three or more characters.

Sources