Integrated nested Laplace approximation for Bayesian nonparametric phylodynamics

摘要: The goal of phylodynamics, an area on the intersection phylogenetics and population genetics, is to reconstruct size dynamics from genetic data. Recently, a series nonparametric Bayesian methods have been proposed for such demographic reconstructions. These rely prior specifications based Gaussian processes proceed by approximating posterior distribution trajectories via Markov chain Monte Carlo (MCMC) methods. In this paper, we adapt integrated nested Laplace approximation (INLA), recently approximate inference latent models, estimation trajectories. We show that when genealogy sampled individuals can be reliably estimated data, INLA enjoys high accuracy replace MCMC entirely. demonstrate significant computational efficiency over state-of-the-art illustrate INLA-based using simulations genealogies hepatitis C human influenza viruses.