Introduction to phylogenetic networks

In the previous chapter we give a brief introduction to phylogenetic trees. Phylogenetic networks provide an alternative to phylogenetic trees and may be more suitable for datasets whose evolution involve significant amounts of reticulate events caused by hybridization, horizontal gene transfer, recombination, gene conversion or gene duplication and loss [56, 61, 89, 201, 219, 231]. Moreover, even for a set of taxa that have evolved according to a tree-based model of evolution, phylogenetic networks can be usefully employed to explicitly represent conflicts in a dataset that may, for example, be due to mechanisms such as incomplete lineage sorting or to inadequacies of an assumed evolutionary model [125]. While rooted phylogenetic networks can, in theory, be used to explicitly describe evolution in the presence of reticulate events, their calculation is difficult and computational methods for doing so have not yet matured into practical and widely used tools [24, 98, 106, 127, 225, 237]. In contrast, there are a number of established tools for computing unrooted phylogenetic networks, which can be used to visualize incompatible evolutionary scenarios in phylogeny and phylogeography [9, 10, 11, 32, 52, 122, 125]. In practice, most currently available algorithms for computing phylogenetic networks are based on combinatorics and this book focuses on such approaches. Some approaches developed within a maximum-parsimony or maximum-likelihood framework can be found, for example, in [59, 106, 141, 142, 143, 228]. In this chapter, we give an introduction to the topic of phylogenetic networks, very briefly describing the fundamental concepts and summarizing some of the most important methods that are available for the computation of phylogenetic networks.