A Note on Evolutionary Rate Estimation in Bayesian Evolutionary Analysis: Focus on Pathogens

Bayesian evolutionary analysis provide a statistically sound and flexible framework for estimation of evolutionary parameters. In this method, posterior estimates of evolutionary rate (μ) are derived by combining evolutionary information in the data with researcher’s prior knowledge about the true value of μ. Nucleotide sequence samples of fast evolving pathogens that are taken at different points in time carry evolutionary information that allow for estimation of evolutionary rates and divergence dates. If the amount of genetic change in the data is proportional to the time elapsed since divergence from the common ancestor, then one can directly estimate the μ from the data. Otherwise, external sources should be used to select the μ value, and use it as a fixed prior in Bayesian evolutionary analysis. This note provides a brief overview on how to assess the adequacy of the evolutionary information in the data and provides some recommendations for obtaining proper evolutionary rate priors from external sources. The recommendations generally highlight the need for the candidate μ prior to be a good representative of the evolutionary rate in the data at hand. This will be achieved by ensuring that the samples that are the source of the candidate μ value have been under relatively similar evolutionary forces as the data at hand. As the evolutionary forces acting on a particular set of samples varies across different study settings and species type, selection of prior for μ should be founded on a thorough understanding of the species under study at biological and social levels.