Measuring Correlation of Evolution Rates Across Multiple Loci: Inference from Large-Scale Sequencing for Infectious Disease Surveillance

• Activity Number: 618
• Type: Contributed
• Date/Time: Thursday, August 7, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Max Tolkoff*+ and Marc Suchard
• Companies: University of California, Los Angeles and University of California, Los Angeles
• Keywords: Phylogenetics ; H5N1 ; Rotavirus ; Evolution Rate

Abstract

Multiple genetic loci may evolve unevenly throughout time, and identifying such heterogeneity across genes from rapidly evolving human pathogens carries implications for disease prevention. We develop a Bayesian phylogenetic model to measure the correlation of evolution rates among multiple genetic loci simultaneously. In brief, the model posits a hierarchical structure across lineage-specific rates of mutation and, in spite of the high- dimensionality of the model, conveniently yields to inference using Markov chain Monte Carlo methods. We apply this novel model to complete genomes from avian influenza H5N1 virus in Southeast Asia and rota virus from a single hospital. Estimates of correlation across influenza genes should help to identify historical reassortment events that give rise to present-day strains. For the rota virus, of the statistically significant across-gene correlation estimates, we find three distinct dependent blocks consisting of two genes and one distinct block with three genes. Two genes in the viral genomic remain independent. Finally, we explore power to detect heterogeneity across genes as functions of evolutionary rate and genome structure.