Phylogenomics of the coconut (Cocos nucifera L.)

Abstract

The coconut palm (Cocos nucifera L.) is a monotypic member of the Cocoseae tribe (subtribe Attaleinae) and its evolutionary history is profoundly intertwined with that of human civilization. It is well adapted to drift-dispersal by oceanic currents, colonizing coastal ecosystems and islands. Both today and in the past, humans have exploited it as a potable source of water, nutritious food, fibre and shelter during their prehistoric voyages of civilization across the Pacific and Indo-Atlantic Oceans. This long-term human interaction and dissemination has altered its phenotype and the lack of a universal domestication trait has obscured the putative wild phenotype and its original geographical location. The main objectives of this phylogenomic study of the coconut are: 1) to determine the centre of coconut domestication, 2) elucidate the geographical origin of the coconut, 3) identify hotspots of genetic diversity, 4) understand migration and gene flow patterns and 5) the impacts of domestication on coconut genome size. Bayesian analysis of population genetic structure was applied to multi-locus microsatellites generated from 1,322 coconut accessions from across the species range. Results strongly suggest that coconuts are differentiated into two genetic populations corresponding to the Indo-Atlantic and Pacific oceanic basins. This pattern suggests independent regions of domestication in these two regions and proposed two centres: island Southeast Asia and the southern margins of the Indian subcontinent. I uncovered evidence for admixtures between these populations consistent with Austronesian trade routes from Southeast Asia to Madagascar and Arab trading along east African coast. To address the overarching objective of the geographical origin of the coconut, I integrated the sub-disciplines of phylogeography, phylogenetics and population genetics to evaluate four criteria: i) ancestral haplotype location, ii) phylogeny and divergence times, iii) coalescence and ancestral reconstruction and iv) genetic diversity. I applied high throughput sequencing technology from chloroplast (14 loci) and nuclear (4 loci) genomes from 118 coconuts across 19 subpopulations representing the species’ distribution. Evaluation of criteria using genomic-scale sequence data, taken together with fossil evidence, suggest that the ancestral geographical origin of the extant coconut is likely in Australasia encompassing Australia, Indonesian Archipelago and Papua New Guinea. The Indo-Atlantic is a hotspot for genetic diversity and a sink population. Migration patterns and gene flow directions were inferred by testing hypotheses of migration models based on geographical and genetic a priori implementing Bayesian coalescent framework and Log Bayes Factors (LBF). For first set of models, LBF indicated that the coconut is not panmictic. The network model showed migration trend from out of Southeast Asia into Oceania consistent with Austronesian migrations. For the second set, bi-directional gene flow model between the Indo-Atlantic and Pacific showed best support. The impact of domestication on genome size and ploidy levels was investigated by flow cytometry technique. Quantifications of genome size of 23 cultivars including Talls, Dwarfs, hybrids and wild-sown coconuts indicate variation. My findings demonstrated that highly domesticated Dwarf types expressed significantly less genome size variation than the Tall types. Ancestral reconstruction of genome sizes amongst Attaleinae show that polyploidy evolved independently at least four times.