Deciphering the process of genetic differentiation of species with insular distributions is relevant for biogeographic and conservation reasons. Despite its importance as old gondwanic islands and part of the Western Indian Ocean biodiversity hotspot, little is known concerning the genetic structure of taxa from the Seychelles islands. We have examined patterns of structure and isolation within Urocotyledon inexpectata (Reptilia: Geckkonidae), an endemic species from this archipelago. Genetic diversity was screened from populations across the archipelago for both mitochondrial and nuclear genes. Gene genealogies and model-based inference were used to explore patterns and timings of isolation between the main lineages. High levels of genetic diversity were found for the mitochondrial and some of the nuclear markers. This species harbours at least two highly differentiated lineages, exclusively distributed across the northern and southern groups of islands. The main split between these was dated back to the Miocene/Late Pliocene, but isolation events throughout the Pliocene and Pleistocene were also inferred. Migration between groups of islands was apparently non-existent, except for one case. The low dispersal capabilities of this species, together with the intrinsic fragmented nature of its geographic distribution seem to have resulted into highly structured populations, despite the cyclic periods of contact between the different island groups. These populations may currently represent more than one species, making U. inexpectata another example of morphologically cryptic lineages with deep genetic divergence within gekkonids. The observed patterns suggest a hypothetical biogeographic scenario (of a main north – south phylogeographic break) for the Seychelles that can be further tested with the exploration of the phylogeographic structure of other Seychellois taxa.
urocotyledon_citb_haplotypes
nexus file of haplotypes of cytochrome-b used to phylogenetic analysis (figure 2). Whole alignments of unphased sequences can be obtained from public genbank database (acession numbers given in the manuscript). Only further sequence alignment files used for analyses are provided here. All "specimen" codes in the files are as given in Appendix S1 of manuscript ("a" and "b" added when needed to represent the different haplotypes of the same individual.
urocotyledon_cmos
fasta file of haplotypes for the fragment of c-mos gene for all individuals (note that no heterozygous individuals were observed, so for this gene both haplotypes for each individual are identical). This is the alignment used for the MJ networks and all further analysis.
urocotyledon_pdc_phased
fasta file of haplotypes from the pdc fragment for all individuals (note that no multiple heterozigous were observed, so all haplotypes - phases - are determined with certainty)
urocotyledon_rag2_phased08_network
fasta file of haplotypes infered for the rag2 dataset for all individuals. Here, positions were phases were resolved with < 0.8 posterior probability are coded with N. This alignment was used for MJ network construction (figure 3), summary statistics, and IMa analysis.
urocotyledon_rag2_phased08_ambiguities
fasta file of haplotypes infered for rag2 dataset for all individuals but where positions whose phases were resolved with < 0.8 posterior probability are coded with IUPAC ambiguity codes. Used for EBSP / species-tree analyses.
urocotyledon_mc1R_phased08_network
fasta file of haplotypes infered for the mc1r dataset for all individuals. Here, positions were phases were resolved with < 0.8 posterior probability are coded with N. This alignment was used for MJ network construction (figure 3), summary statistics and IMa analysis.
urocotyledon_mc1R_phased08_ambiguities
fasta file of haplotypes infered for mc1r dataset for all individuals but where positions whose phases were resolved with < 0.8 posterior probability are coded with IUPAC ambiguity codes. Used for EBSP / species-tree analyses.
urocotyledon_reln_phased08_network
fasta file of haplotypes infered for the reln dataset for all individuals. Here, positions were phases were resolved with < 0.8 posterior probability are coded with N. This alignment was used for MJ network construction (figure 3) and summary statistics.
urocotyledon_reln_phases08_ambiguities
fasta file of haplotypes infered for reln dataset for all individuals but where positions whose phases were resolved with < 0.8 posterior probability are coded with IUPAC ambiguity codes. Used for EBSP.
urocotyledon_reln_phased08_bestguesses
fasta file of haplotypes infered for reln dataset for all individuals, where positions whose phases were resolved with < 0.8 posterior probability are nevertheless coded with the respective best guesses for each base. (as in this dataset the different runs of PHASE (Stephens et al., 2001 - to be fund on manuscript references) were inconsistent at some positions we used results of the run with the highest likelihood score, whose seed (and command line) is given bellow. The known haplotypes used (uro_reln.known) were those read directly from sequences heterozigous for indels, and that can be found on genbank - accession numbers given in the paper). [phase comand line: ./PHASE -MS -p0.8 -q0.8 -d1 -s48993 -kuro_reln.known uro_reln.inp uro_reln.out 10000 1 100]
urocotyledon_reln_imgc_Cw03
fasta file resulting from the trimming of the file "urocotyledon_reln_phased08_network.fas" to the maximum block free of evidence of recombination, using the program IMgc (Woerner et al., 2007 - reference to be found in the ms) with a chromosome copy weight of 0.3 (the value chosen as it maximized the sequence lenght keeping all needed populations represented). Used for IMa and species-tree analyses (that assume no recombination within each sequence fragment)
urocotyledon_reln_bestguesses_imgc_Cw03
fasta file resulting from the trimming of the file "urocotyledon_reln_phased08_bestguesses.fas" to the maximum block free of evidence of recombination, using the program IMgc (Woerner et al., 2007 - reference to be found in the ms) with a chromosome copy weight of 0.3 (the value chosen as it maximized the sequence lenght keeping all needed populations represented). Used for IMa and species-tree analyses (that assume no recombination within each sequence fragment)