15 - From sequence reads to evolutionary inferences

Summary

Introduction

The history of molecular systematics can be caricatured as one of ever-increasing depth of sequence data, analysed by ever more complex models. In this respect, sequence data from whole genomes are the ultimate source of molecular markers that can act as characters for phylogenetic or population genetic analysis. While complete genomes in the strictest sense are only available for very few species, and fragmentary genome assemblies that capture the entire genome, but in many pieces, are also fairly restricted in scope beyond the prokaryotes, this is changing rapidly. More-or-less shallow genomic data, for example from EST sequencing projects, high-throughput transcriptome sequencing or some other kind of reduced-representation sequencing (see review by Davey et al. 2011) are now becoming widespread and of increasing utility in systematics and other areas of evolutionary biology. Studies using these kinds of data to reconstruct relationships between species have become known as ‘phylogenomics’, although the original usage of the term referred to using phylogenetic approaches to infer gene function (Eisen 1998), and the other parts of the research programme proposed under this name (Eisen and Fraser 2003) have been subsumed into the broader study of comparative and evolutionary genomics. Moreover, the term ‘phylogenomics’ has, perhaps, become over-extended, as datasets that claim this title vary in size and can be as few as 11 markers (Horvath et al. 2008) or as little as 30 kb of sequence data (Wiegmann et al. 2011), and in eukaryotic organisms, the ‘genomes’ in question are very often organelle (mitochondrial or chloroplast) genome sequences. Sequence data from whole genomes have the potential to be a rich source of molecular phylogenetic markers for any systematic question, but there are two areas in which large-scale, highly multi-locus data appear most valuable – occupying the two extremes of the range of timescales over which inference about evolutionary history is made.

Genome-scale data promise the ability to resolve ancient divergences, and in particular, fairly rapid (at least in geological terms) ancient radiations that have been difficult to reliably reconstruct with more limited molecular datasets. In this context, phylogenomic data have been applied to a wide taxonomic range of phylogenetic questions. Early usage of whole-genome data was in prokaryote systematics (e.g. Daubin et al. 2002).