Evolutionary similarity among genes

• 成果类型: Article
• 署名作者: Suchard, MA; Weiss, RE; Sinsheimer, JS; Dorman, KS; Patel, M; McCabe, ERB
• 署名单位: University of California System; University of California Los Angeles; University of California Los Angeles Medical Center; David Geffen School of Medicine at UCLA; University of California System; University of California Los Angeles; University of California System; University of California Los Angeles; University of California Los Angeles Medical Center; David Geffen School of Medicine at UCLA; Iowa State University; Iowa State University; University of Toronto; Hospital for Sick Children (SickKids); University of California System; University of California Los Angeles; University of California Los Angeles Medical Center; David Geffen School of Medicine at UCLA
• 刊物名称: JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
• ISSN/ISSBN: 0162-1459
• DOI: 10.1198/016214503000000558
• 发表日期: 2003
• 页码: 653-662
• 关键词: chain monte-carlo; bayesian phylogenetic inference; dna-sequences; mitochondrial-dna; sex-determination; markov-chains; P-values; models; sry; trees

摘要

An evolutionary history of a set of organisms is a family tree, or topology, with branches of various lengths between vertices that describe how closely the organisms are related to each other. We consider the K evolutionary histories of K genes from a set of N organisms. Evolutionary similarity (ES) occurs when the branching patterns and relative branch lengths in the K evolutionary histories of the genes are the same or nearly the same across the set of organisms. Evolutionary similarity indicates similarity of evolutionary pressures acting on these genes. Current likelihood approaches identify ES conditional on a given topology. For a variety of reasons, different genes may support different topologies when fit independently. We use Bayesian models and reversible-jump Markov chain Monte Carlo to jointly infer topology and branch lengths for multiple genes simultaneously. We test for ES using Bayes factors, conditionally on a consistent topology over the multiple genes, where the topology is either known or unknown. We relax the single topology assumption by employing a dissimilarity measure between evolutionary histories and testing for ES using both prior and posterior predictive p values. We apply our methodology to three genes (DAX1, SOX9, and SRY) believed to be involved in sex determination in primates. We find support in the data for ES between DAX1 and SRY, but not SOX9. These results are consistent with the hypothesized biological roles of these genes.