Phylotranscriptomics and genome-size evidence clarify the Taiwanese Cirsium japonicum complex and delimit C. brevicaule and allied East Asian thistles.

Résumé

BACKGROUND: Rapid diversification in island floras often creates taxonomic uncertainty, particularly for morphologically variable plant complexes. The complex, a widespread and medicinally important group in East Asia, exemplifies this challenge, with unclear species boundaries and conflicting varietal definitions. This is especially true in Taiwan, where multiple endemic forms co-occur. Despite its ecological and pharmacological significance, due to morphological overlap, fragmented distributions, and the absence of comprehensive phylogenomic analysis, the evolutionary relationships and lineage delimitations within this complex remain unresolved. In this study, we analyzed phylotranscriptomic data from 37 thistle accessions comprising the complex and the allied group, constructed multigene coalescent species trees, and integrated morphometrics, genome size mapping, demographic history, and distribution modeling.RESULTS: Three monophyletic subsections (, , and ; all within sect) diverged 1.30–1.02 million years ago (Mya) (95% highest posterior density (HPD): 1.62–0.71 Mya) during early Quaternary glaciations. Within subsect. , five lineages emerged: Japanese var. and four Taiwanese varieties that form two sister pairs (var. vs. var. and var. vs. var. ). In subsect. , two distinct species endemic to the Ryukyu Islands, and , exist across the Miyako Strait. The ancestral genome size was estimated at ~1.3 pg and underwent several independent reductions (e.g., var. , 1.01 pg) and expansions (e.g., , 1.93 pg) without chromosomal changes. Skyline plots indicate a late Pleistocene bottleneck and Holocene rebound in var. , whereas var. remained stable, matching historical habitat suitability.CONCLUSIONS: We combined phylotranscriptomic trees, genome-size trajectories, and demographic models to resolve East Asian into two Ryukyu species (, ) and a five-lineage complex spanning Taiwan to Japan. Revised diagnoses confirm that true is confined to the central Ryukyus, while Taiwanese records reflect the complex, including var. , , , and the Hengchun endemic var. , which meets IUCN vulnerability criteria due to restricted range and medicinal harvest. We further clarify the placement of , link floral-color polymorphism in var. to anthocyanin expression and pollinator preference, and show that Quaternary glaciations, island fragmentation, and 2C shifts shaped diversification, whereas late Pleistocene bottlenecks and Holocene expansions structured populations. Together, this synthesis refines taxonomy, informs conservation, and supports sustainable use.SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12870-026-08097-6.