Aim: Oceanic islands have played an important role in our understanding of the diversification of organisms, and phylogenetic estimates have been used in this context to investigate the origin of island diversity and its relationship to the continent. Using a typical orchid genus rich in island endemics and with widespread continental relatives, we aim to compare alternative hypotheses of diversification with a focus on island endemism.

Location: Tropical Africa and the Gulf of Guinea Islands of São Tomé & Príncipe, Central Africa.

Taxon: Tridactyle genus (Orchidaceae).

Methods: We used genome skimming to sequence the whole chloroplast genome and nuclear ribosomal genes from 157 individuals of 34 Tridactyle species and 15 individuals of 12 other orchid genera (outgroups) to infer a time-calibrated phylogenetic tree of the genus Tridactyle. We also used multiple statistical methods to infer the geographic ranges of the ancestral nodes from the estimated phylogeny. Alternative hypotheses for the origins of endemism on the islands of São Tomé and Príncipe were investigated based on the biogeographic reconstruction of the genus.

Results: The estimated phylogeny of Tridactyle and reconstruction of geographic ranges for the ancestral nodes suggested a general history of allopatric speciation for the genus, in particular via colonisation of the islands of the Gulf of Guinea that induced a long period of geographic isolation. The most parsimonious hypothesis to explain island endemism in Tridactyle involved 6 independent colonisations of the islands from the continent.

Main conclusions: In contrast to other cases of oceanic island endemism that involved adaptive radiation on an island or within an archipelago, endemism in Tridactyle is better explained by multiple colonisation events from the continent to São Tomé and Príncipe, with subsequent divergences due to geographic isolation but only one potential instance of further diversification on the islands.

Genome skimming on 157 samples from 34 species of Tridactyle (out of 47), sampled across the African continent and São Tomé and Príncipe. Most of the samples were obtained from fieldwork and the shadehouse network from Central Africa (Stévart, 2003), from São Tomé, Príncipe, Equatorial Guinea, Guinea Conakry, Cameroon, Gabon, Rwanda, Kenya, Tanzania and South Africa. Plants from Cameroon and Gabon were grown in shadehouses monitored by our teams in Yaoundé (for those collected in Cameroon) and in Libreville, Tchimbélé and Sibang (for those collected in Gabon). Herbarium specimens from the Kew Botanical Garden collection were sampled, as well as living plants from the Botanical Garden of Meise. 

We conducted separate phylogenetic analyses on two separate datasets: one containing all samples (157 Tridactyle and 16 outgroups)  and the other including a single sample per species (61 samples).

Alignement of 173 sequences (157 Tridactyle (34 out of 47), 4 other Vandeae, 9 other Epidendroideae and 3 Orchidoideae). It allowed us to obtain a plastid alignment of 162,994 bp, partitionned with IQ-Tree (Chernomor et al., 2016; Nguyen et al., 2015) (i.e. 173sequencesChloroPart.nex), a ribosomal genome of 7212 bp (i.e. 173sequencesRibo.nex), and a combined dataset of both, partitionned with IQ-Tree (i.e. 173partitionned.nex)

Alignement of 61 sequences (49 Tridactyle, 3 other Vandeae, 7 other Epidendroideae and 2 Orchidoideae). It allowed us to obtain a plastid alignment of 162,994 bp, partitionned with IQ-Tree (i.e. 61SequencesChloroOK.nex), a ribosomal genome of 7212 bp (i.e. 61sequencesRiboOK.nex), and a combined dataset of both, partitionned with IQ-Tree (i.e. 61Seq.nex)

Chernomor, O., von Haeseler, A., & Minh, B. Q. (2016). Terrace Aware Data Structure for Phylogenomic Inference from Supermatrices. Systematic Biology, 65(6), 997‑1008. https://doi.org/10.1093/sysbio/syw037
Nguyen, L.-T., Schmidt, H. A., von Haeseler, A., & Minh, B. Q. (2015). IQ-TREE : A Fast and Effective Stochastic Algorithm for Estimating Maximum-Likelihood Phylogenies. Molecular Biology and Evolution, 32(1), 268‑274. https://doi.org/10.1093/molbev/msu300