Missing data, incomplete taxa, and phylogenetic accuracy

Area of Science:

• Phylogenetics
• Computational Biology
• Evolutionary Biology

Background:

• Missing data is a significant challenge in phylogenetic analysis, especially for fossil taxa.
• Incomplete taxa are often assumed to be problematic due to the amount of missing data they contain.
• Previous studies indicate that including incomplete taxa can reduce phylogenetic accuracy and resolution.

Purpose of the Study:

• To investigate the underlying mechanisms causing incomplete taxa to negatively impact phylogenetic analyses.
• To determine whether the proportion of missing data or the number of complete characters is the primary cause of reduced accuracy.
• To re-evaluate common practices for handling incomplete taxa in phylogenetic reconstruction.

Main Methods:

• Phylogenetic analysis using simulation studies.
• Comparison of tree reconstruction accuracy with varying levels of data completeness.
• Analysis of the impact of missing data versus limited complete characters on phylogenetic outcomes.

Main Results:

• Reduced phylogenetic accuracy in analyses with incomplete taxa is caused by a lack of sufficient complete characters, not the proportion of missing data.
• The amount or proportion of missing data is not the direct cause of the 'missing data problem' for incomplete taxa.
• Incomplete taxa can be accurately placed phylogenetically, even with high levels of missing data.

Conclusions:

• Taxon exclusion based solely on completeness levels may be misguided.
• A revised understanding of the missing data problem can lead to more effective strategies for incorporating incomplete taxa.
• The findings challenge assumptions underlying some phylogenetic methods, such as the supertree approach, and highlight potential for accurate placement of incomplete taxa.