Phylogenomics of the Olea europaea complex using 15 whole genomes supports recurrent genetic admixture together with differentiation into seven subspecies

Abstract

Background: The last taxonomic account of Olea recognises six subspecies within Olea europaea L., including the Mediterranean olive tree (subsp. europaea) and five other subspecies (laperrinei, guanchica, maroccana, cerasiformis, and cuspidata) distributed across the Old World, including Macaronesian islands. The evolutionary history of this monophyletic group (O. europaea complex) has revealed a reticulated scenario involving hybridization and polyploidization events, leading to the presence of a polyploid series associated with the subspecies. However, how the polyploids originated, and how the different subspecies contributed to the domestication of the cultivated olive are questions still debated. Tracing the recent evolution and genetic diversification of the species is key for the management and preservation of its genetic resources. To study the recent history of the O. europaea complex, we compared newly sequenced and available genomes for 27 individuals representing the six subspecies.

Results: Our results show discordance between current subspecies distributions and phylogenomic patterns, which support intricate biogeographic patterns. The subspecies guanchica, restricted to the Canary Islands, is closely related to subsp. europaea, and shows a high genetic diversity. The subsp. laperrinei, restricted now to high mountains of the Sahara desert, and the Canarian subsp. guanchica contributed to the formation of the allotetraploid subsp. cerasiformis (Madeira islands) and the allohexaploid subsp. maroccana (western Sahara region). Our phylogenomic data support the recognition of one more taxon (subsp. ferruginea) for the Asian populations, which is clearly segregated from the African subsp. cuspidata.

Conclusions: In sum, the O. europaea complex underwent several processes of hybridization, polyploidy, and geographical isolation resulting in seven independent lineages with certain morphological traits recognised into subspecies.

Keywords: Macaronesian islands; Olive (Olea europaea); Phylogenomics; Polyploidy; Subspecies.

Fig. 1

Olea europaea sampling and SNV densities (SNVs/Kb). a Geographical distribution of the six O. europaea subspecies (modified from [8]). Samples sequenced in this project are marked with green stars and other whole-genome sequences with red stars. Colours show the distribution of each subspecies. b SNVs/Kb for homozygous (darker colours) and heterozygous (lighter striped colours) SNVs in the nuclear genome. c SNV densities for the plastid (dark colour) and mitochondrial (light striped colour) genomes

Fig. 2

Density plot for the relative coverage of alternative alleles in heterozygous sites per individual. For all cases, we only plotted data corresponding to the 23 pseudo-chromosomes. Red line indicates the position of the peak consistent with the ploidy

Fig. 3

The SNV similarity between diploids and polyploids and the possible origin of the polyploids. a Heatmap showing the percentage of the SNV positions of the polyploids (cerasiformis and maroccana) shared with the diploid subspecies. Only the exclusive SNVs of the diploid subspecies were analysed. b Map depicting the possible origin of the polyploids. Approximate distribution of subsp. laperrinei in yellow and subsp. maroccana in pink. The empty circles show the distribution of subspp. cerasiformis (purple) and guanchica (blue). The dotted line rectangle represents the presumed ancestral distribution of subsp. guanchica. The distribution of the subspecies was taken from [22]

Fig. 4

Evolutionary relationships among the individuals of the O. europaea complex. Maximum likelihood species tree for the plastid SNVs (a), mitochondrial SNVs (b), and the nuclear SNVs (c). The geographical location of the accession and the plastid haplotype are indicated. Only bootstrap values below 100% are shown. d Bayesian clustering for the nuclear SNV data estimated in Structure v2.3 for the O. europaea complex. Structure bar plot shows the genetic clusters differentiated by colour. e Heatmap showing the D-statistic and its p value. Red colour indicates higher D-statistics, and more saturated colours indicate greater significance

Fig. 5

SplitsTree derived from nuclear SNVs. a Complete plot. b Zoom in on the region in the rectangle. The neighbor-net method is used here to explore data conflict and not to estimate phylogeny