"Bayesian inference of mixed Gaussian phylogenetic models"Brahmantio, BayuContinuous trait evolution is commonly modelled using stochastic differential equations (SDEs) to represent deterministic change of trait through time, while incorporating noises that represent different unobservable evolutionary pressures. Two of the most popular classes of SDEs are Brownian motion (BM) and Ornstein-Uhlenbeck (OU) process, which fall under a larger family of models called GLInv that has a Gaussian transition probability with expectation that is linear with respect to ancestral value and variance that is invariant with respect to it. This framework enables multiple different GLInv models under a single phylogenetic tree to capture diversity of traits from different species on the tips. In this work, a Bayesian scheme is implemented as an extension of the maximum likelihood method to include uncertainties in the parameter estimates and prior knowledge that are more biologically relevant. The method is written as an R package that utilizes Monte Carlo inference methods based on importance sampling to retrieve posterior quantities and to evaluate and compare models. |
« back