Ecological speciation in palms on an oceanic island

The origin of species diversity has challenged biologists for over two centuries. Charles Darwin recognised that allopatry, species divergence resulting from geographical isolation, is a driving force in speciation, but he also thought that populations could diverge into separate species in the absence of geographical isolation, a mechanism now called sympatric speciation. During the last decade, mathematical models have shown that sympatric speciation is theoretically possible, but extremely few examples have been documented in nature. Early this year, Savolainen and colleagues (some of the applicants of this proposal) provided complete evidence for sympatric speciation in a case study of two species of palm, Howea forsteriana and H. belmoreana, endemic of the remote Lord Howe Island (LHI) ('sympatric speciation in palms on an oceanic island', Nature 441: 210-213, 2006). Here we propose to take our research to a much deeper level. Our goal is to document the genomic architecture and genetic basis for species divergence in our case study of palms, thus bringing new types of data to test alternatives to divergence in sympatry and disentangle the mechanisms of ecological speciation that must have operated in this unique LHI system. LHI is a minute subtropical island of less than 12 km2, situated 580 km off the eastern coast of Australia. Its flora comprises 241 vascular plant species of which 105 are endemic. Geographic isolation is not realistically possible on LHI and thus it is an ideal site on which to test the four criteria for sympatric speciation: 1) species sympatry, 2) sister relationships, 3) reproductive isolation, and 4) that an earlier allopatric phase is highly unlikely. Numerous plant genera, like the palms Howea, are represented by more than one endemic species on the island, which may well be products of sympatric speciation. In our previous experiments of the Howea palms, we found that the genetic signature of species divergence indicated that only four loci differ more strongly between the two species than expected under neutrality and these are those most likely to be linked to genes subject to divergent selection during sympatric speciation. To characterise the genomic and genetic architecture underlying the ecological speciation of Howea in this unique island system, we will conduct six activities: a) Using a more comprehensive AFLP genome scan, identify in Howea the most divergent marker loci and so regions of the genome that may be under selection b) Using these divergent loci as probes on a BAC library, characterise portions of the kentia palm's genome that may contain genes involved in species divergence (testing Gavrilets' prediction of only a few loci being involved in speciation in Howea) c) Using the BACs sequences of the kentia palm as template, identify in the curly palm's genome homologous sequences and open reading frames (ORFs), which may contain the signature of positive selection d) Determine whether the divergent genomic segments are concentrated versus widely distributed in the genomes of the two palm species (testing Felsenstein's classic prediction of speciation being facilitated by low recombination). e) Using coalescence models and multiple loci, evaluate whether gene flow has been interrupted / and estimate the extent of disruption / during species divergence in Howea. f) In comparison with an EST library for the oil palm, provide possible functions for any ORFs found under selection.