Phylogenomics Reveals Three Sources of Adaptive Variation during a Rapid Radiation

Abstract

Speciation events often occur in rapid bursts of diversification, but the ecological and genetic factors that promote these radiations are still much debated. Using whole transcriptomes from all 13 species in the ecologically and reproductively diverse wild tomato clade (Solanum sect. Lycopersicon), we infer the species phylogeny and patterns of genetic diversity in this group. Despite widespread phylogenetic discordance due to the sorting of ancestral variation, we date the origin of this radiation to approximately 2.5 million years ago and find evidence for at least three sources of adaptive genetic variation that fuel diversification. First, we detect introgression both historically between early-branching lineages and recently between individual populations, at specific loci whose functions indicate likely adaptive benefits. Second, we find evidence of lineage-specific de novo evolution for many genes, including loci involved in the production of red fruit color. Finally, using a "PhyloGWAS" approach, we detect environment-specific sorting of ancestral variation among populations that come from different species but share common environmental conditions. Estimated across the whole clade, small but substantial and approximately equal fractions of the euchromatic portion of the genome are inferred to contribute to each of these three sources of adaptive genetic variation. These results indicate that multiple genetic sources can promote rapid diversification and speciation in response to new ecological opportunity, in agreement with our emerging phylogenomic understanding of the complexity of both ancient and recent species radiations.

Fig 1. Geographic distribution and ecological diversity of sampled populations of wild tomato.

(A) Wild tomato species inhabit diverse ecological zones (shaded regions) along the western coast of South America and the Galápagos Islands. For each sample location, labels indicate species and accession number, and symbols denote major phylogenetic groupings (circle = Esculentum, triangle = Arcanum, square = Peruvianum, star = Hirsutum, oval = outgroup; base map modified from original from http://www.freevectormaps.com). High variation of (B) altitude, (C) mean annual temperature (D), and annual precipitation across the habitat range of wild tomato species (data from http://www.worldclim.org; plotted using GRASS GIS http://grass.osgeo.org/).

Fig 2. The phylogeny of Solanum sect. Lycopersicon.

(A) A whole-transcriptome concatenated molecular clock phylogeny with section Lycopersicoides as the outgroup. Branch colors indicate the four major subgroups (labels on right). Pie charts on each node indicate majority rule extended bipartition support scores (out of 100) using trees from 100-kb genomic windows. All nodes are supported by 100 bootstrap replicates, except “*” denotes bootstrap support score of 68. (B) A “cloudogram” of 2,745 trees (grey) inferred from nonoverlapping 100-kb genomic windows (see also S2 Table). For contrast, the consensus phylogeny is shown in black.

Fig 3. Low variation and pervasive sorting of ancestral heterozygosity in the wild tomato clade.

(A) The fraction of heterozygous sites within accessions and (B) the proportion of heterozygous sites in each accession that are sorting in other species.

Fig 4. Multiple modes of molecular adaptation.

(A) The D-statistic for testing introgression for 100-kb windows on the short arm of chromosome 1 using the tree shown. Shaded regions indicate windows where introgression is significantly detected (p ≤ 1.45 × 10⁻⁴; see also S1 Data 1.5 and S5A Fig). (B) Alleles are shown indicating an example of a lineage-specific substitution for the red-fruited Esculentum group, with colored dots indicating approximate fruit colors for each accession. Genes directly involved in the carotenoid biosynthesis pathway were found to have several nonsynonymous (^N) and synonymous (^S) substitutions specific to the red-fruited lineage. (C) Environment-specific alleles found in groups of accessions from different phylogenetic groups but under common environmental stressors for consistent (sun) or variable (sun/cloud) climate (top) or the presence (filled symbol) or absence (empty symbol) of heavy metals in the environment (bottom).