6.07 - Phylogenetic Analyses

Abstract

The construction of phylogenetic trees by which one can study the evolutionary history of species has become a ubiquitous part of biological analyses. Historically, phylogenetic trees were constructed by comparing the morphological features of the organisms studied. Nowadays, as a result of sequencing efforts, phylogenetic studies are mainly performed on the DNA or protein sequences of organisms. From a set of input sequences, a multiple sequence alignment (MSA) is constructed that serves as standard input on which phylogenetic tree building is based. MSA may delineate sequence regions that undergo rapid changes during evolution (and therefore might be less functionally constrained) or reveal residues showing evidence of being shaped by natural selection. Phylogenetic analysis is an important tool for researchers looking for structural and functional annotation of their input sequence set, where the sequences can be closely related, for instance, representing a conserved homologous family, or may represent a more widely divergent group. Given the large variety of phylogenetic methods, researchers using phylogenetic analyses in their research should understand the underlying technical aspects and assumptions of these methods, in order to be able to choose the right type of method for different situations. Sequence-based phylogenetic tree-building methods can be broadly categorized into two main groups: distance-based and character-based methods. Distance-based methods are computationally less demanding than character-based methods, which makes these methods better suited for reconstructing phylogenetic trees for large numbers of sequences. On the other hand, character-based methods, although much slower than distance-based methods, are in most cases more accurate and generally yield phylogenetic trees that reflect the evolutionary events more optimally. The most recent character-based methods for estimating phylogenetic trees are based upon Bayesian statistics. Bayesian methods have been proposed as a relatively fast and efficient means to provide the user with the possibility to incorporate complex and appropriate evolutionary models for the case at hand into the phylogenetic tree-building process.