Analysis of the distribution of bootstrap tree lengths using the maximum parsimony method

Area of Science:

• Evolutionary Biology
• Bioinformatics
• Computational Biology

Background:

• Bootstrap analysis is crucial for estimating confidence intervals in phylogenetic trees.
• Interpretation of bootstrap probability values lacks a standardized consensus in evolutionary studies.

Purpose of the Study:

• Investigate the bootstrap method for phylogenetic analyses.
• Explore the utility of bootstrap tree distribution characteristics for interpreting evolutionary data.

Main Methods:

• Applied bootstrapped maximum parsimony (MP) analyses to nine small subunit ribosomal DNA (SSU rDNA) datasets.
• Utilized both unweighted and weighted sequence positions in the MP analyses.
• Analyzed the lengths (parsimony steps) and distribution of bootstrap trees.

Main Results:

• The shape and mean of bootstrap tree distributions provide insights into evolutionary signal.
• Complex phylogenies with multifurcations may show significant differences between bootstrap distribution means and best tree lengths.
• Weighting sequence positions enhances bootstrap support at internal nodes but can lead to conflicting groupings.
• A correlation between tree topology used for weighting and bootstrap consensus tree topology influences results.

Conclusions:

• Bootstrap tree distribution characteristics, such as shape and mean, offer valuable insights into phylogenetic signal.
• Weighting schemes can be selected based on bootstrap tree distribution properties like skewness.
• Understanding bootstrap distribution is key to robust phylogenetic inference.