In a recent Opinion article published in Nature Reviews Molecular Cell Biology (Computational morphodynamics of plants: integrating development over space and time. Nature Rev. Mol. Cell Biol. 12, 265–273 (2011))1, Roeder and collaborators integrate some of the recent work in the emerging field of plant computational morphodynamics, focusing on models based on cellular live-imaging data. One of the examples these authors review is that of Arabidopsis thaliana root epidermal patterning, for which they discuss two mathematical models, namely the WEREWOLF (WER) self-activation model, which we authored2, and the mutual support model3. They rule out the WER self-activation model, based on experimental data showing that expression of the MYB transcription factor WER is the same in wild-type cells and wer loss-of-function mutant cells. Moreover, Roeder et al. did not consider the results and discussions presented in a more recent version of the WER self-activation model4 and in a further theoretical paper5.

We argue that the WER self-activation model4 — which assumes that WER can positively regulate itself, either directly or through the action of other genes2,4 — cannot be ruled out based only on the expression of WER in wer mutants. A similar situation was encountered for the AGAMOUS self-regulatory loop in flower morphogenesis6. In the case of A. thaliana root epidermal patterning, recent experimental data show that WER expression is downregulated by the Leu-rich repeat receptor-like kinase SCRAMBLED (SCM)7,8,9 and that SCM expression is, in turn, downregulated by WER8,9, giving rise to a positive feedback loop. This results in the indirect and local self-activation of WER and stabilizes SCM-mediated position signalling. The presence of WER in wer mutants, which was used to discount our model, indeed seems to contradict these data. However, both pieces of evidence can be reconciled if WER is also regulated by other factors, and thus its promoter can still be activated in wer mutants. Furthermore, additional evidence in support of WER being self-activated is that the closely related MYB gene MYB23 — which can substitute for the function of WER10 — participates in a positive feedback loop with itself. This suggests that the observed cell pattern can still occur in wer mutants, owing to the self-maintained expression levels of MYB23 (Ref. 10).

This correspondence may be important for the interpretation of experimental data and modelling studies in other cases, given that the coexistence of various apparently redundant circuits seems to be relevant for the robustness of morphogenetic systems5. In the particular case of A. thaliana epidermal cell patterning, we argue that the WER self-activation model cannot be ruled out on the basis of the evidence currently available, as published data may still support it. We believe that other circuits, such as the mutual support model, are also consistent with the available data and are possibly reinforcing each other, conferring robustness to the overall spatiotemporal cellular patterns5,11.