The Positive Selection Estimation And Statistic Test For Two Kinds Of Codon Substitution Models

Phylogenetics is an important research field of molecular evolutionary biology, involvingbioinformatics, mathematics, biostatistics, computer science and so on. In recent years, thecomputational molecular evolution based on phylogenetics develops rapidly and has been widelyapplied in the fields of evolutionary genetics, ecology and biological experiment science such asvirology and developmental biology. We usually introduce Markov chain to construct substitutionmodels when we study the protein-coding DNA sequences. Substitution model is the importantapproach in study of molecular evolution, so the study for the substitution model is meaningful.In this paper, the positive selection estimation and statistic test for two kinds of codonsubstitution models are studied. In chapter2, by using the CUSP program, we calculate the value ofsynonymous codon usage bias; Based on synonymous codon usage bias, a new substitution model isestablished, and we use the maximum likelihood approach to estimate parameters and detect whetherthe species under the positive selection. In chapter3, two substitution models are presented based ontwo kinds of transversion and the discrete distribution for selection variable. The traditionaltransversion is divided into two categories: A Tor G Cand A Cor T G, andfrom this the first codon substitution model is established. Based on the first model, we establish thesecond codon substitution model in which the selection variable is discretely distributed. We alsocompare the suggested models with existing codon models by applying them to real data. In chapter4,the statistical standard for the reconstruction of phylogenetic trees, the test for reliability ofphylogenetic trees, and the methods of test for selection of substitution models are discussed.