genotype
transgene
A betalain construct repaints whatever organs its promoter reaches. Which organs those are is read out of the network, not painted on by hand.
| whorl | organ | n | A | B | C | E | WUS |
|---|
Pick a genotype. It is run through the published Boolean network for floral organ identity, whorl by whorl, and the flower those attractors specify is drawn. New to this? The walkthrough teaches the whole subject from scratch in fifteen steps.
A betalain construct repaints whatever organs its promoter reaches. Which organs those are is read out of the network, not painted on by hand.
| whorl | organ | n | A | B | C | E | WUS |
|---|
The network is the Chaos et al. 2006 correction of the Espinosa-Soto, Padilla-Longoria & Álvarez-Buylla floral organ specification model, transcribed verbatim from the encoding distributed with the CANA project. It has fifteen nodes, two of which are held constitutively on, and it settles into exactly ten fixed points: four inflorescence states plus sepal, petal, stamen and carpel. The build checks that our transcription reproduces all ten published attractors bit for bit, and that the classic mutants come out right.
Determinacy is not bolted on. WUSCHEL maintains itself and is switched off only by AGAMOUS acting with SEPALLATA, so when C function is missing the centre of the flower never terminates and the whorl series simply repeats inward. The doubled flower is an output of the published rules, not a special case.
The network says what a cell can become, never where it is. Position enters as a per-whorl starting state — A function seeded in whorls one and two, the UFO cofactor in two and three, the organising centre in three and four. That pre-pattern is our choice; it reproduces the textbook mutants, which is evidence for it, not proof of it.
Everything about shape is invention. Real organ form comes out of growth models we are not running, so outlines here are parameterised caricatures tuned to be recognisable and to interpolate. Organ numbers are the canonical wild-type counts held fixed per whorl position; real flowers vary. Pigment colours are qualitative stand-ins, not spectra. And SUPERMAN, which bounds the B domain and matters for real doubled flowers, is not a node in this network at all, so it cannot be modelled here.
Chaos A, Aldana M, Espinosa-Soto C, García Ponce de León B, Garay Arroyo A, Álvarez-Buylla ER. From genes to flower patterns and evolution: dynamic models of gene regulatory networks. J Plant Growth Regul 25(4):278–289, 2006. Network encoding via the CANA project. RUBY: He et al., Hortic Res 7:152, 2020.