HyPhy: Hypothesis Testing Using Phylogenies

Abstract

6.1 Introduction The field of molecular evolution, though wide-reaching in its breadth, can be split into two types of investigations: studies of phylogeny and studies of the molecular evolutionary process. Of course, each of these two categories en-compasses many different types of questions, and many investigations require studies of both phylogeny and evolutionary process, but the proposed binary classification is a useful construct. Software for molecular evolution is focused disproportionately on problems relating to phylogenetic reconstruction, with a number of outstanding comprehensive packages from which to choose. On the other hand, software for addressing questions of the molecular evolutionary process tends to be found in stand-alone programs that answer only one or two quite specific problems. The HyPhy system, available for download from www.hyphy.org, was designed to provide a unified platform for carrying out likelihood-based analyses on molecular evolutionary data sets, the emphasis of analyses being the molecular evolutionary process; that is, studies of rates and patterns of the evolution of molecular sequences. HyPhy consists of three major components: a high-level programming lan-guage designed to facilitate the rapid implementation of new statistical meth-ods for molecular evolutionary analysis; a collection of prewritten analyses for carrying out widely used molecular evolutionary methods; and a graphical user interface that allows users to quickly and interactively analyze data sets of aligned sequences using evolutionary models and statistical methods that they design using the software system. This chapter is intended to provide an overview of the key elements of each of the three system components, in-cluding both specific details of the basic functionality as well as a conceptual description of the potential uses of the software. The nature of the package prevents the creation of an exhaustive " cookbook " of available methods. In-stead, we hope to provide a collection of fundamental tools and concepts that allow users to begin using HyPhy to carry out both existing and new methods of data analysis. The second of the three enumerated HyPhy components was a collection of prewritten " standard " analyses. Since this section of the software is essen-tially just a collection of prepackaged analyses, we will not devote much time to a detailed discussion of it. However, we choose to describe it first in this chapter to illustrate the types of analyses that HyPhy has been designed to address. In Figure 6.1, we show the initial Standard Analyses menu invoked by Analyses:Standard Analyses... (note the use of Small Caps to indi-cate menu items, with submenus or selections separated by a colon). Each of the nine major headings includes a collection of routines that can be selected by the user. For example, the Positive Selection menu item expands to offer five different analyses relating to the task of identifying nucleotide sites undergoing positive selection. A total of 35 batch files are included in the col-lection, and most of these files include a variety of options enabling users to select items such as evolutionary models or topology search methods. Topics include molecular clock tests, positive selection analyses, phylogenetic recon-struction, and model comparison procedures. The authors frequently add new standard analyses to the package. HyPhy includes the ability to perform Web updates, which ensures that the distribution is kept up-to-date. Fig. 6.1. HyPhy Standard Analyses menu (Mac OS X).