Discoveries made in experimental models provide insight into the mechanisms working in other organisms. The models developped by the plant communities include Arabidopsis, Physcomitrella patens, Antirrhinum and petunia. The release of the Arabidopsis genome in 2000 granted this, already popular, experimental model a further advantage over other equally important species. In this issue of Nature Plants, Li et al. report the genome of the, much overlooked, Antirrhinum majus (snapdragon), removing its previous weaknesses and reinforcing its status as a research model.

The Antirrhinum genus used to be a classic model for flowering plants before the epoch of Arabidopsis. In fact, its history of being a research workhorse is very long, tracing back to the earliest inheritance studies of Darwin and Mendel in the nineteenth century. In the twentieth century, Erwin Baur established Antirrhinum as a model because it had specific advantages that made it a tractable experimental subject1.

Native to the western Mediterranean, Antirrhinum has a very high level of natural phenotypic diversity, as recognized by the earliest geneticists. For example, the 18 species that make up the Antirrhinum genus exhibit a very wide range of interspecific variations in flower colour and pattern, fragrance, flowering time, and much else. Genetic instability widely exists in the Antirrhinum genus, constantly generating new variations; an intriguing phenomenon in itself. Their relatively small size and hardiness make it easy to cultivate and keep Antirrhinum on a large scale. These advantages, together with a small genome (only four times that of Arabidopsis), a short generation time of about three months, and the ease of performing self- and cross-pollination, made it particularly attractive for early experimental geneticists.

Two main collections of Antirrhinum mutants were established early on. One consists of more than 450 mutants, originally generated and collected by Baur, now available from the IPK-Gatersleben in Germany. The other is comprised of more than 300 mutants, generated by Rosemary Carpenter and colleagues at the John Innes Center in the UK. Besides these research materials, standard molecular biological tools, such as genomic libraries and yeast two-hybrid libraries, as well as genetic tools such as linkage maps, have also been developed. However, genetic manipulations of Antirrhinum proved more difficult than in Arabidopsis.

The pioneering use of Antirrhinum by Baur and colleagues in the first several decades of the twentieth century led to a series of discoveries in plant genetics, such as the first confirmation of genetic linkage, cytoplasmic inheritance and genes responsible for flower colour and morphology2. In the middle of the twentieth century, work on unstable mutations of Antirrhinum isolated the first autonomous transposon from a plant, Tam13. Subsequently, more transposons were found and used to facilitate gene isolation through transposon-tagging systems.

The most important contribution of Antirrhinum research was in the area of developmental genetics. The first plant MADS-box gene, DEFICIENS (the orthologue of the Arabidopsis APETALA3), was cloned in Antirrhinum4, as well as the floral-organ identity gene GLOBOSA, and floral-meristem identity genes LEAFY and APETALA1. These homeotic mutants (together with related mutations in Arabidopsis), laid the foundations for the ABC model of flower patterning5. Also, zygomorphy has been intensively examined using Antirrhinum, and transcription factors such as CYCLOIDEA and RADIALIS were found to control its asymmetric growth6. The founding member of the S-LOCUS F-BOX genes controlling self-incompatibility was also found in Antirrhinum7. Moreover, Antirrhinum research has shed light on leaf development, anthocyanin biosynthesis, pigmentation and scent emission, and how these factors affect pollinator attraction, speciation and adaptive evolution.

Antirrhinum continues to attract researchers today, especially for investigations that are not possible with other model systems, such as perennial habit and flower colour evolution. It can be expected that the new, near-complete chromosome-scale genome of A. majus presented by Li et al. will make experiments with this model plant far more convenient. For example, a recent work reported in Science using this reference genome identified a selection target underlying the evolution of flower colour pattern8.

No single species can represent all plant diversity. Recent developments in sequencing technologies, particularly long-read sequencing, provide opportunities to generate highly accurate and continuous reference genomes for many species at low costs, reducing the demand for simple genetics in experimental systems. Many species will become tractable for genetic and genomic studies and we will be better equipped to study lineage- or species-specific biological phenomena using new models. Investigations of previously overlooked experimental plants will thrive again, thanks to the new sequence resource. For functional studies, ease of transformation will always be a limitation, but breakthroughs can be expected.

However, it is far too soon to announce the end of model system research and trumpet the rise of non-model plants. If nothing else, research on model species promotes scientific cooperation and communication, tying communities together by promoting conversations and exchanges of experimental materials and ideas. Overlooked models — which Antirrhinum has some justification to consider itself — will provide stepping stones to the investigation of a more diverse range of species and questions. A clear picture of plant biology will only be secured by exploring a larger diversity of the plant kingdom: model, non-model and neglected-model plants alike.