Haplowebs as a graphical tool for delimiting species: a revival of Doyle's "field for recombination" approach and its application to the coral genus Pocillopora in Clipperton
1 Courant Research Center "Geobiology", University of Göttingen, Goldschmidtstr. 3, 37077 Göttingen, Germany
2 GENOSCOPE, Centre National de Séquençage, 2 rue Gaston Crémieux, CP5706, 91057 Evry Cedex, France
3 UMR UPMC-CNRS-MNHN-IRD 7138, Département Systématique et Évolution, Muséum National d'Histoire Naturelle, Case Postale 26, 57 rue Cuvier, 75231 Paris Cedex 05, France
4 URBO, Department of Biology, University of Namur, Rue de Bruxelles 61, 5000 Namur, Belgium
BMC Evolutionary Biology 2010, 10:372 doi:10.1186/1471-2148-10-372Published: 30 November 2010
Usual methods for inferring species boundaries from molecular sequence data rely either on gene trees or on population genetic analyses. Another way of delimiting species, based on a view of species as "fields for recombination" (FFRs) characterized by mutual allelic exclusivity, was suggested in 1995 by Doyle. Here we propose to use haplowebs (haplotype networks with additional connections between haplotypes found co-occurring in heterozygous individuals) to visualize and delineate single-locus FFRs (sl-FFRs). Furthermore, we introduce a method to quantify the reliability of putative species boundaries according to the number of independent markers that support them, and illustrate this approach with a case study of taxonomically difficult corals of the genus Pocillopora collected around Clipperton Island (far eastern Pacific).
One haploweb built from intron sequences of the ATP synthase β subunit gene revealed the presence of two sl-FFRs among our 74 coral samples, whereas a second one built from ITS sequences turned out to be composed of four sl-FFRs. As a third independent marker, we performed a combined analysis of two regions of the mitochondrial genome: since haplowebs are not suited to analyze non-recombining markers, individuals were sorted into four haplogroups according to their mitochondrial sequences. Among all possible bipartitions of our set of samples, thirteen were supported by at least one molecular dataset, none by two and only one by all three datasets: this congruent pattern obtained from independent nuclear and mitochondrial markers indicates that two species of Pocillopora are present in Clipperton.
Our approach builds on Doyle's method and extends it by introducing an intuitive, user-friendly graphical representation and by proposing a conceptual framework to analyze and quantify the congruence between sl-FFRs obtained from several independent markers. Like delineation methods based on population-level statistical approaches, our method can distinguish closely-related species that have not yet reached reciprocal monophyly at most or all of their loci; like tree-based approaches, it can yield meaningful conclusions using a number of independent markers as low as three. Future efforts will aim to develop programs that speed up the construction of haplowebs from FASTA sequence alignments and help perform the congruence analysis outlined in this article.