Phylogenomics and classification of Cactaceae based on hundreds of nuclear genes

Abstract

Phylogenetic classification based on evolutionary relationships is the standard approach in systematics, but Cactaceae has posed significant challenges due to the signature of its rapid radiation: low sequence divergence hindering phylogenetic resolution and enormous species diversity hindering attempts to adequately reflect phylogenetic diversity. Previous classifications mostly relied on joint assessment of multiple phylogenetic studies and/or intuition on morphological evolution, lacking comprehensive genomic analysis. Here, we propose a revised phylogenetic classification of Cactaceae, based on the Angiosperms353 set of phylogenomic markers, including 170 species, covering close to 90% of genera and common segregates. Coalescent-based gene tree-species tree reconciliation reveals a well-resolved phylogenetic backbone, mostly congruent with a previous plastid DNA-based summary phylogeny. Some unresolved areas surrounding the subfamilies Cactoideae and Pereskioideae remain however, where gene concordance analyses reveal complex evolutionary histories. We formally recognize the four traditional subfamilies Pereskioideae, Opuntioideae, Maihuenioideae, and Cactoideae, plus Blossfeldioideae and Leuenbergerioideae. Four subfamilies are monogeneric, while within Opuntioideae, we recognize three tribes (Opuntieae, Cylindropuntieae, and Pterocacteae) and within Cactoideae, we recognize eight tribes: Lymanbensonieae, Copiapoeae, Cacteae (incl. subtribes Echinocactinae, Ferocactinae, Cactinae), Phyllocacteae (incl. subtribes Corryocactinae, newly recognized Leptocereinae, Hylocereinae, Echinocereinae), Fraileae, Rhipsalideae, Notocacteae, and Cereeae (incl. subtribes Aylosterinae, Rebutiinae, Gymnocalyciinae, Cereinae, newly recognized Reicheocactinae, Trichocereinae). Our completely revised classification for all clades includes full generic synonymy accepting 155 genera. Overall, the phylogenetic structure of Cactaceae mirrors the angiosperm-wide pattern of enigmatic, species-poor lineages dispersed amongst clades that are orders of magnitude more species rich, revealing a mix of nested, 'explosive' radiations and orphan lineages.

Supplementary information: The online version contains supplementary material available at 10.1007/s00606-025-01948-z.

Keywords: Angiosperms353; Cactaceae; Multispecies coalescent; Paraphyly; Phylogenetic classification; Phylogenetic conflict; Phylogenomics; Rapid radiation.

Fig. 1

Cacti are attractive and sometimes landscape-dominating elements in both North (a Anza Borrego desert, USA, California, with Cylindropuntia acanthocarpa and Ferocactus acanthodes) and South America (b Chile, with Eulychnia acida and Cumulopuntia sphaerica), ubiquitous elements in Botanical Gardens in glasshouse (c Sukkulenten-Sammlung Zürich) and outdoor plantings (d Desert Garden at the Huntington Botanical Garden, San Marino, California, USA) around the world, but also fascinating collectibles and study objects for amateur collections (e a collection of Weingartia species), but also invasive neophytes in suitable frost-free climates worldwide (f Opuntia ficus-indica naturalized in Spain, Andalusia, Almería, Níjar, Las Negras). Credits: a-c, e Urs Eggli, d CC-BY Flickr cultivar413, f Felix Merklinger

Fig. 2

a ASTRAL tree (pro parte) based on the QC data set. Branch support is indicated above branches as local posterior probability. Suprageneric names are marked: subfamilies are colour-coded to the right, while tribes and subtribes are labelled at their corresponding stem lineage. Higher taxa for which monophyly could not be confirmed are marked with an asterisk (*). Tips are labelled according to their accepted genus, as listed in Table 3. Alternative genus names are provided after a dash for species that represent lineages commonly treated as separate genera, while "ss" is indicated if a species represents the type lineage of a genus that is commonly split. Genus names in brackets indicate taxa that were previously placed in a larger genus complex. P and SRR numbers correspond to newly obtained and public data, respectively, and are linked to Online Resource 1 for accession details. The scale bar represents coalescence units. b ASTRAL tree (pro parte) based on the QC data set. Annotation as in (a). c ASTRAL tree (pro parte) based on the QC data set. Annotation as in Fig. 2a

Fig. 2

a ASTRAL tree (pro parte) based on the QC data set. Branch support is indicated above branches as local posterior probability. Suprageneric names are marked: subfamilies are colour-coded to the right, while tribes and subtribes are labelled at their corresponding stem lineage. Higher taxa for which monophyly could not be confirmed are marked with an asterisk (*). Tips are labelled according to their accepted genus, as listed in Table 3. Alternative genus names are provided after a dash for species that represent lineages commonly treated as separate genera, while "ss" is indicated if a species represents the type lineage of a genus that is commonly split. Genus names in brackets indicate taxa that were previously placed in a larger genus complex. P and SRR numbers correspond to newly obtained and public data, respectively, and are linked to Online Resource 1 for accession details. The scale bar represents coalescence units. b ASTRAL tree (pro parte) based on the QC data set. Annotation as in (a). c ASTRAL tree (pro parte) based on the QC data set. Annotation as in Fig. 2a

Fig. 2

a ASTRAL tree (pro parte) based on the QC data set. Branch support is indicated above branches as local posterior probability. Suprageneric names are marked: subfamilies are colour-coded to the right, while tribes and subtribes are labelled at their corresponding stem lineage. Higher taxa for which monophyly could not be confirmed are marked with an asterisk (*). Tips are labelled according to their accepted genus, as listed in Table 3. Alternative genus names are provided after a dash for species that represent lineages commonly treated as separate genera, while "ss" is indicated if a species represents the type lineage of a genus that is commonly split. Genus names in brackets indicate taxa that were previously placed in a larger genus complex. P and SRR numbers correspond to newly obtained and public data, respectively, and are linked to Online Resource 1 for accession details. The scale bar represents coalescence units. b ASTRAL tree (pro parte) based on the QC data set. Annotation as in (a). c ASTRAL tree (pro parte) based on the QC data set. Annotation as in Fig. 2a

Fig. 3

Multidimensional Scaling (MDS) analysis of genetic distinctness across (a) all Cactaceae with subfamilies indicated and b Cactoideae with tribes indicated, based on concatenated alignments of the QC dataset. Colours indicate subfamilies and polygons indicate non-monogeneric tribes

Fig. 4

BUCKy analyses of main and alternative topologies. a Primary concordance tree with internal nodes proportional to coalescence units and sample gene concordance factors indicated. b-f) Subtrees with sample gene concordance factors indicated for five selected, alternative topologies that deviate from the primary concordance tree: b, c surrounding Pereskia (where ss indicates the sensu stricto lineage of Pereskioideae) d, e surrounding Opuntioideae, f surrounding Cactoideae. Tip names indicate higher taxa, and accession numbers indicate species listed in Online Resource 1