Detection of Implausible Phylogenetic Inferences Using Posterior Predictive Assessment of Model Fit

• DOI: 10.1093/sysbio/syu002
• Publication Date: 2014-01-12T01:29:06Z
• Authors: Jeremy M. Brown
• Keywords: Evolution, Molecular; 0303 health sciences; 03 medical and health sciences; Bayes Theorem; Computer Simulation; Models, Theoretical; Phylogeny

Systematic phylogenetic error caused by the simplifying assumptions made in models of molecular evolution may be impossible to avoid entirely when attempting model across massive, diverse data sets. However, not all deficiencies inference result unreliable estimates. The field phylogenetics lacks a direct method identify cases where specification adversely affects inferences. Posterior predictive simulation is flexible and intuitive approach for assessing goodness-of-fit assumed priors Bayesian analysis. Here, I propose new test statistics use posterior assessment fit. These compare inferences from sets original data. A study demonstrates utility these statistics. tests reject plausibility inferred tree lengths or topologies more often data/model combinations produce biased also apply this exemplar empirical sets, highlighting value novel assessments. [Bayesian; Markov chain Monte Carlo; fit; phylogenetic; distribution; sequence evolution; simulation.