STATISTICALMETHODS IN PHYLOGENETIC AND EVOLUTIONARY INFERENCES

Molecular instruments are the most accurate methods in organisms ’ identification and characterization. Biologists are often involved in studies where information about ge-netic features of organisms is fundamental and where the main goal is to identify re-lationships among individuals. In this framework, it is very important to know and apply the most robust approaches to infer correctly these relationships, allowing the right conclusions about phylogeny. Together with the establishment of newmolecular techniques, scientists developed a large number of different methods to compare individuals basing on differences among them. In his book, Joseph Felsenstein (2004) provides a very good and deep explanation of these techniques, together with a detailed description of phylogenetic tree building methods. This review will introduce the reader to the most used statistical methods in phy-logenetic analyses, the Maximum Likelihood and the Bayesian approaches, considering for simplicity only analyses regarding DNA sequences. In the first section we will introduce the concept of the Maximum Likelihood and we will apply it the phylogenetic trees, including models of molecular evolution and explaining how to draw a tree. In the next section we will move on Bayesian statistics applied to phylogenetic inferences: we will introduce the a posteriori concept and we will describe how to create tree topology using the Metropolis Markov chains Mon-tecarlo algoriths. Finally, we will show several examples of studies conducted to infer about the evolution of pathogens