Reconstructing the origins and the biogeography of species' genomes in the highly reticulate allopolyploid-rich model grass genus Brachypodium using minimum evolution, coalescence and maximum likelihood approaches.

The identification of homeologous genomes and the biogeographical analyses of highly reticulate allopolyploid-rich groups face the challenge of incorrectly inferring the genomic origins and the biogeographical patterns of the polyploids from unreliable strictly bifurcating trees. Here we reconstruct a plausible evolutionary scenario of the diverging and merging genomes inherited by the diploid and allopolyploid species and cytotypes of the model grass genus Brachypodium. We have identified the ancestral Brachypodium genomes and inferred the paleogeographical ranges for potential hybridization events that originated its allopolyploid taxa. We also constructed a comprehensive phylogeny of Brachypodium from five nuclear and plastid genes using Species Tree Minimum Evolution allele grafting and Species Network analysis. The divergence ages of the lineages were estimated from a consensus maximum clade credibility tree using fossil calibrations, whereas ages of origin of the diploid and allopolyploid species were inferred from coalescence Bayesian methods. The biogeographical events of the genomes were reconstructed using a stratified Dispersal-Extinction-Colonization model with three temporal windows. Our combined Minimum Evolution-coalescence-Bayesian approach allowed us to infer the origins and the identities of the homeologous genomes of the Brachypodium allopolyploids, matching the expected ploidy levels of the hybrids. To date, the current extant progenitor genomes (species) are only known for B. hybridum. Putative ancestral homeologous genome have been inherited by B. mexicanum, ancestral and recent genomes by B. boissieri, and only recently evolved genomes by B. retusum and the core perennial clade allopolyploids (B. phoenicoides, B. pinnatum 4x, B. rupestre 4x). We dissected the complex spatio-temporal evolution of ancestral and recent genomes and have detected successive splitting, dispersal and merging events for dysploid homeologous genomes in diverse geographical scenarios that have led to the current extant taxa. Our data support Mid-Miocene splits of the Holarctic ancestral genomes that preceded the Late Miocene origins of Brachypodium ancestors of the modern diploid species. Successive divergences of the annual B. stacei and B. distachyon diploid genomes were implied to have occurred in the Mediterranean region during the Late Miocene-Pliocene. By contrast, a profusion of splits, range expansions and different genome mergings were inferred for the perennial diploid genomes in the Mediterranean and Eurasian regions, with sporadic colonizations and further mergings in other continents during the Quaternary. A reliable biogeographical scenario was obtained for the Brachypodium genomes and allopolyploids where homeologous genomes split from their respective diploid counterpart lineages in the same ancestral areas, showing similar or distinct dispersals. By contrast, the allopolyploid taxa remained in the same ancestral ranges after hybridization and genome doubling events. Our approach should have utility in deciphering the genomic composition and the historical biogeography of other allopolyploid-rich organismal groups, which are predominant in eukaryotes.