Estimating errors and confidence intervals for branch lengths in phylogenetic trees by a bootstrap approach

Area of Science:

• Evolutionary Biology
• Computational Biology
• Phylogenetics

Background:

• Estimating branch-length errors and confidence intervals in phylogenetic trees is crucial for understanding evolutionary relationships.
• Existing methods often assume equal substitution rates among lineages, which may not reflect biological reality.
• Accurate assessment of internodal distances is essential for robust phylogenetic inference.

Purpose of the Study:

• To propose a novel bootstrap-based method for estimating branch-length errors and confidence intervals in phylogenetic trees.
• To develop a method applicable to phylogenies where equal rates of substitution among lineages do not necessarily hold.
• To enable testing whether estimated internodal distances are significantly greater than zero.

Main Methods:

• A bootstrap procedure is employed for error estimation and confidence interval calculation.
• The method is compatible with any estimator of genetic distances.
• It can be used in conjunction with any tree reconstruction procedure based on distance matrices.

Main Results:

• The proposed method provides a robust way to estimate branch-length errors and confidence intervals.
• It effectively tests the significance of internodal distances, even with unequal substitution rates.
• The method's applicability is demonstrated using the phylogenetic tree of four hominoid species.

Conclusions:

• The developed bootstrap-based method offers a flexible and reliable approach to phylogenetic error estimation.
• It addresses limitations of methods assuming equal substitution rates, enhancing phylogenetic accuracy.
• The technique is versatile, supporting various genetic distance estimators and tree reconstruction algorithms, and is scalable to any number of species.