Comparative chloroplast genomics of Cardueae (Carduoideae, Asteraceae): the first tribe-wide analysis

The tribe Cardueae (Carduoideae, Asteraceae) comprises approximately 2,600 thistle-like species across temperate regions worldwide, yet chloroplast (cp.) genomic resources remain limited, with published cp. genomes for only 18 genera. Here, we present the first comprehensive comparative cp. genomics study of Cardueae, achieved by sequencing and/or de novo assembly of 54 species from 37 genera. The de novo assembled cp. genomes were combined with 18 previously reported cp. genomes retrieved from NCBI. In total, cp. genomes of 72 species representing 51 genera across all 12 subtribes of Cardueae were compared. Chloroplast genomes ranged from 151,098 bp to 153,475 bp, exhibiting a typical quadripartite structure with highly conserved gene content (113 unique genes) and organization. Only Serratula tinctoria showed a trnL-UAA inversion, and Mantisalca salmantica exhibited matK pseudogenization. However, assembly artifacts cannot be ruled out for these two species. Inverted repeat boundaries, amino acid frequency, and codon usage showed high similarity among species. Codon usage favored A/T-ending codons, and substitution analysis indicated that transversions were more frequent than transitions, with a transition/transversion ratio ranging from 0.54 to 0.96. Phylogenetic analyses using maximum likelihood and Bayesian inference strongly supported the monophyly of nine of the twelve recognized subtribes, whereas Onopordinae and Saussureinae were polyphyletic. The placement of Synurus deltoides (Onopordinae) within Saussureinae revealed cytonuclear discordance that contrasts with nuclear phylogenetic studies. Similarly, two species of Saussureinae, Polytaxis lehmannii and Saussurea chabyoungsanica, were grouped with Arctiinae, rendering Saussureinae non-monophyletic. Dipterocominae, represented by a single species (Dipterocome pusilla), requires additional sampling to assess monophyly. This study provides the most comprehensive cp. genomic framework for Cardueae to date, establishing essential resources for phylogenetic and conservation research. The detection of cytonuclear discordance highlights the importance of integrating nuclear and cp. data for accurate taxonomic inference in rapidly radiating lineages.