An integrative approach for phylogenetic inference

Molecular evolution is the key to explain the divergence of species and the origin of life on earth. The main task in the study of molecular evolution is the reconstruction of evolutionary trees from sequences data of current species. This is a huge, complicated and highly contentious field and in the literature a large number of phylogenetic reconstruction methods have been proposed to address such a problem. This thesis first researches into the mathematical and statistical foundations of the subject and discusses the limitations of existing methods. All the methods reconstruct the phylogeny optimising an explicit criterion. These criteria attempt to explain the real evolutionary process by simplified assumptions under a certain number of models and parameters, all of which are more or less biased in practice. Therefore, all current available phylogenetic reconstruction methods have limitations. This has become a serious issue for many important real-life applications which demand very accurate results. Therefore, it is desirable to have an integrative approach which can effectively incorporate different methods, useful information and analytical results from various sources during dynamic computation and to allow biologists to interact with the computation to control the quality of the analysis at various processing stages. In this way, we may achieve biologically higher accuracy and reliability for large and complex real-life problems. This thesis presents a novel integrative and interactive distributed high performance computing system for phylogenetic analysis. It is a cross-disciplinary project and involves sophisticated research, design and development of algorithms and systems: ( l) the design of new algorithms which are able to effectively integrate heterogeneous information from various computational and experimental studies, (2) the design of new service-oriented high performance computing technologies which enable efficient dynamic integration and interaction, and (3) the development of a new distributed high performance computing platform which allows biologists to interact effectively with the computation to control the quality of the analysis in solving their large and complex problems. This work also provides a preliminary algorithmic framework for the coestimation of multiple sequence alignment and phylogenetic reconstruction