Testing Phylogenetic Trees with Least Squares Methods
Author Information
Author(s): Czarna Aleksandra, Sanjuán Rafael, González-Candelas Fernando, Wróbel Borys
Primary Institution: Institute of Oceanology, Polish Academy of Sciences
Hypothesis
Can the least squares approach effectively construct confidence sets of phylogenetic trees from biological data?
Conclusion
The WLS method is a computationally efficient alternative to GLS for constructing confidence sets of trees, especially when assessing phylogenetic signals.
Supporting Evidence
- The WLS method did not negatively affect the accuracy and reliability of the test.
- The GLS method often included all possible trees in the confidence set despite strong phylogenetic signals.
- The study used a well-known data set of mammalian mitochondrial proteins for analysis.
Takeaway
This study looks at how to check if a tree showing relationships between species is correct, using a method that makes calculations easier.
Methodology
The study compares the generalized least squares (GLS) and weighted least squares (WLS) methods for constructing confidence sets of phylogenetic trees using biological sequence data.
Potential Biases
Potential biases arise from model misspecification and the assumptions of normality in distance measures.
Limitations
The GLS method often fails to calculate the test statistic for many data sets, and both methods may struggle with closely related taxa.
Participant Demographics
The study analyzed sequences from six mammalian species: cow, harbor seal, human, mouse, opossum, and rabbit.
Statistical Information
P-Value
0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website