Speciation, hybridization and extinction are major factors in plant evolution, but only the timing and nature of the first two can be detected, often with precision, using molecular data. Conversely, extinction has proved almost undetectable, except using fossils. Birth–death models, which estimate extinction rates across a phylogeny, appear to be unreliable. However, hybrid speciation presents an opportunity to detect extinct or “ghost” species, which might leave their genomes, or part thereof, in their hybrid species offspring. This provides a possible pathway for detecting extinct taxa and placing them onto molecular phylogenies.

The alpine tetraploid Allium tetraploideum (2n = 4x = 32), one of the five known members in the subgenus Cyathophora. In combination with multiple methods, we expected to reveal whether extinct lineage(s) might have participated the origin of this tetraploid Allium. This is a good choice to study dynamics of species diversification in the Qinghai–Tibet Plateau when climatic oscillations occurred.

Recently, a paper titled ‘Evolutionary origin of a tetraploid Allium species on the Qinghai–Tibet Plateau’ was published online in Molecular Ecology, a well-known journal in the field of Ecology, by Institute of Innovation Ecology & School of Life Sciences, Lanzhou University. In this study, by combining ecological adaptation, phylogenetic genomics, genome-level paraplegic homologous differentiation, simplified genome and multilocus population genetics data, and fluorescence in situ hybridization (GISH and FISH) analyses, the authors have systematically elucidated the adaptation and complex origin of this tetraploid Allium. The results showed that the tetraploid Allium is an obvious allotetraploid, and its origin is very complex, involving both polyploidy speciation and tetraploid level homoploid hybridization speciation. Its ancestors include two extant closely related diploid species (Allium cyathophorum and A. farreri), one extinct diploid and two extinct tetraploid ancestor lineages. 

The authors first showed that the alpine tetraploid Allium tetraploideum (2n = 4x = 32) has distinct adaptions to high altitudes in seed weight, stomata size and seedling growth advantage compared to its two diploid relatives (Figure 1). Second, Allium phylogenomics using chloroplast genomes and transcriptomes revealed a close relationship between A. tetraploideum and A. cyathophorum + A. farreri, with morphological similarity among them. However, three significant Ks peaks were identified within the paired homologues of A. tetraploideum, of which one corresponded closely to the homologous divergence between A. cyathophorum and A. farreri, and the older peaks indicates an extra diploid lineage has participated the origin of A. tetraploideum, because the homologous divergence peak between extinct lineage and A. cyathophorum + A. farreri has not been observed among peaks of the extant diploid pairs, suggesting the existence of an extinct lineage. In addition, phylogenetic analyses on various nuclear loci, involving multiple individuals and 10 cloned sequences per individual, repeatedly showed admixture of some sequences from A. tetraploideum with all those of A. cyaphoporum, whereas other A. tetraploideum sequences formed one or two clades elsewhere in the tree, most often sister to A. farreri, or occasionally closer to A. spicatum. Such topological structure can be well explained by using the participation of an extinct lineage to the origin of A. tetraploideum.

Figure 1. Left: The significantly larger stomata and seeds of A. tetraploideum compared to A. cyathophorum and A. farreri. Right: Phylogenomics of Allium subgenus Cyathophora using chloroplast genomes and transcriptomes.

To further test the hybridization-extinction hypothesis for the origin of A. tetraploideum, using population-scaled SLAF data, the authors revealed a clearly different genetic complement for A. tetraploideum when compared with A. cyathophorum and A. farreri, and the coalescence simulations indicated that the simplest pathway for deriving A. tetraploideum from three ancestors would be for an earlier diverging, extinct diploid to have first formed allopolyploids independently with each of A. farreri and A. cyathophorum, and those two allotetraploids would then have crossed to produce A. tetraploideum by homoploid hybridization speciation (Figure 2). Moreover, two tetraploid ancestors were also extinct. Finally, GISH analysis hybridizing gDNAs of the four diploid relatives onto the A. tetraploideum chromosomes, indicated that its 32 chromosomes were divided into six that hybridized with all the four diploids, 14 that hybridized with A. cyathophorum and/or A. farreri, two that hybridized only with A. spicatum, and 10 that hybridized with no extant diploids, respectively (Figure 2), suggesting these 10 chromosomes would have contributed by the extinct ancestor. Towards this results, at least three extinct lineages have been detected compared to the five extant species in the small subgenus Cyathophora. This suggests that many extinct taxa would have existed in the evolution of species diversification in the QTP and the extinct groups would also contribute much the local diversification.

Figure 2. Left: The best-fitting evolutionary model among A. cyathophorum (CC), A. farreri (FF) and A. tetraploideum (CFGG) using fastsimcoal2, with the participation of a ghost lineage (GG). The inferred split ages are shown in circles in units of millions of years before the present. The red and green labels indicate the extant A. cyathophorum and A. farreri, and blue label and red “×” indicate the extinct lineages. Right: Four-fluorochrome GISH with gDNAs of the four extant diploid species (2n = 2x = 16) probed onto the chromosomes of A. tetraploideum, showing 22 chromosomes stained: chromosomes 1–6 (pale-rose) = four diploid species; chromosomes 7–10 (yellow) = A. cyathophorum + A. farreri; chromosomes 11–16 (red) = A. cyathophorum; chromosomes 17–20 (green) = A. farreri and chromosomes 21–22 (greenish-yellow) = A. spicatum，23-32 blue probably derived from the extinct ancestor.

Minjie Li and Zeyu Zheng contribute equally as first authors. Prof. Jianquan Liu is the corresponding author. This study was supported by Strategic Priority Research Program of Chinese Academy of Sciences and the National Natural Science Foundation of China.

Paper link：https://onlinelibrary.wiley.com/doi/10.1111/mec.16168