A Phylogenetic Latent Liability Model for Assessing Correlations in Phenotype Evolution

• Activity Number: 639
• Type: Topic Contributed
• Date/Time: Thursday, August 8, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Gabriela Cybis*+ and Janet Sinsheimer and Philippe Lemey and Marc A. Suchard
• Companies: UCLA and UCLA and KU Leuven and UCLA
• Keywords: Bayesian phylogenetics ; Brownian diffusion ; Threshold model ; Antimicrobial resistance

Abstract

The latent liability model is used in statistical genetics for traits with discrete outcome determined by an underlying unobserved continuous factor. In an extension of this model to phylogenetics, the underlying continuous factor (latent liability) undergoes Brownian diffusion along the tree. At the tips, the discrete observable trait is defined depending on the position of the latent liability relative to a set of thresholds. By extending this non-Markovian model to the multivariate case, we can use it to make inference about the correlation structure of many characters. A unique feature of the model is that, because of the underlying continuous variable, it can also be used to assess correlations between discrete and continuous trait evolution. We implement this model in the context of Bayesian phylogenetics and present an application to antimicrobial resistance data, where we assess evolutionary correlations between resistance to different drugs.