Gene Sequence Analysis Using Multinomial Model and Generalized Fiducial Inference

• Activity Number: 276
• Type: Topic Contributed
• Date/Time: Tuesday, July 31, 2012 : 8:30 AM to 10:20 AM
• Sponsor: General Methodology
• Author(s): Wen Jenny Shi*+ and Corbin D. Jones and Jan Hannig
• Companies: The University of North Carolina at Chapel Hill and The University of North Carolina at Chapel Hill and The University of North Carolina at Chapel Hill
• Keywords: multinomial ; Dirichlet mixtrue ; Dempster-Schafer models ; MCMC ; DNA sequence ; generalized fiducial inference

Abstract

Analysis on DNA sequence at nucleotide level is crucial for understanding genomes. We construct a Dirichlet mixture model to describe a given gene sequence and propose an effective Markov chain Monte Carlo method that enhanced with a bi-clustering simulation scheme which allows parallel computing. For multinomial model selection, we consider a fiducial method and a Bayesian approach, where one is equivalent to the simplex Dempster-Schafer model (DSM) proposed by A. P. Dempster in 1966, and the other is the Dirichlet-DSM, an alternative approach developed by C. Liu in 2009. We discuss the relationship among the two DSM's, give a comprehensive comparison, and provide some theoretical results. Our simulation method and two multinomial model selection methods are applied to the DNA sequence of Coxsackievirus B.