A fundamental question in biology is how constraints drive phenotypic changes and the diversification of life. While domesticated species are extraordinary models for tracking phenotypic changes in response to directional selection, we know little about the role of these constraints on crop domestication, nor how artificial selection can escape them.
Crop yields experienced spectacular growth during the Green Revolution thanks to plant breeding and agronomic development. Yet, crop improvement has not seen comparable advances for increased productivity in the last decades. Artificial selection over preexisting natural variation might have reached a dead end, and the environmental impacts of intensified agronomy are unsustainable in the long term. Thus, plant breeding may have to look for novel strategies to ensure food security and environmental sustainability for a steadily increasing human world population, including the creation of ideotypes adapted to low-input agricultural management. While new biotechnological tools producing ‘super-domesticates’ may offer qualitative advancement in breeding for higher productivity, plant breeding approaches rarely account for ecophysiological and biophysical constraints - linked to resource capture, use and partitioning - that can strongly limit the improvement of the basic functions of plants: growth, reproduction and survival. These constraints arise from trade-offs (i.e. the impossibility to simultaneously optimize two conflicting functions) that take place at different organizational levels: cell, organ, organism and crop (monoculture and mixture). These constraints are considered as obstacles for the improvement of target crop traits by plant breeders but their quantification remains scarce. Interestingly, several generic, cross-species, ecophysiological and biophysical constraints have been identified in theoretical and empirical ecology, with an acceleration in the last two decades with the rising of trait-based ecology. The CONSTRAINTS project highlighted the persistence of these constraints, found across wild species, within and between domesticated species. We have also identified candidate genes associated to these constraints, which represents a possibility to overcome them in future tradeoff-free improvement approaches.
