Variance to mean ratio, R(t), for poisson processes on phylogenetic trees

Area of Science:

• Molecular Evolution
• Population Genetics
• Bioinformatics

Background:

• The ratio of variance to mean (R(t)) is used to test the neutral theory of molecular evolution and the molecular clock hypothesis.
• Previous studies reported R(t) values exceeding 1, questioning the neutral theory.
• Existing estimators of R(t) assume specific distributions and means.

Purpose of the Study:

• To investigate the accuracy of proposed R(t) estimators.
• To re-evaluate R(t) values in light of accurate statistical distributions.
• To differentiate between lineage effects and violations of the neutral theory.

Main Methods:

• Monte Carlo simulations were used to determine the true means and distributions of R(t) estimators.
• Analysis was performed on both star and general phylogenetic trees.
• Well-known gene sequences were reanalyzed using corrected statistical methods.

Main Results:

• Previously suggested means and distributions for R(t) estimators were found to be highly inaccurate.
• Kimura's estimators are unsatisfactory for statistical testing of R(t) in star phylogenies.
• Bulmer's correction factor was confirmed as accurate for star phylogenies.
• For nonstar phylogenies, R(t) values, though >1, fell within confidence limits under Poisson process models.

Conclusions:

• R(t) estimators require accurate statistical distributions for valid testing.
• Lineage effects, not necessarily neutral theory violations, can explain high R(t) values.
• Distinguishing lineage effects from molecular clock deviations is crucial for understanding molecular evolution.