Our paper published in Nature Ecology and Evolution is available via go.nature.com/2Je7Pyj

Can artificial selection produce domesticates with features absent in wild species? Do domesticates come from specific or from random linages of the tree of life? Some five years ago we began comtemplating this and other questions while drafting a project proposal (and while enjoying reading this blog post). From many possible questions, we thought that these two would already be well addressed in the literature. Suprisingly, this was not the case when we considered domestication at the global scale. Asking this question through the lens of phylogenetic comparative methods made the story most interesting and unexplored. And we got the funding!

The botanist in the kitchen. A most interesting blog on science and food plants. Credit: Catherine Preston & Janne Osnas.

Answering questions at a global scale implied little (though some) experimental work, and lots of interaction with colleagues around the world. Other labs studying crops with a similar mindset rapidly joined the endevaour and were eager to share datasets and outcomes of ongoing projects. Critically, the TRY and PanTHERIA initiatives supplied most of the data for wild species, which were key to establish comparative backgrounds for our domesticates. That work was supplemented with long hours foraging the primary literature and assembling phylogenies. In this collaborative way we put together the largest database on the taxonomic identity and phenotypic profiles, for a few key functional traits, of crops and livestock species and of their wild relatives. We were ready to satiate our curiosity.

Sampling and phenotyping crop plants. A: at PhenoArch-INRA (Montpellier, France). B: at Universidad Rey Juan Carlos (Madrid, Spain). C: at University of Pittsburgh (Pittsburgh, USA). D: researchers from the University of Sheffield (UK) sampling wild wheat in Turkey. Credit: the authors. 

This paper is the first outcome of that project. It nicely shows that crops and livestock are mostly a subset of wild species when it comes to traits such as adult and offspring sizes, or metabolic rates. Thus, even if some crops and domestic mammals might look peculiar, most of that peculiarity was already available in nature. Our results may help overturn common views that domesticates are products that evolve under different rules and should thus be subject of separate disciplines. We also show that crops and livestock differ in a fundamental trait: phylogenetic breadth. While livestock come from a few, clustered lineages, crops come from multiple and diverse lineages, with only modest phylogenetic structure. The fact that the few livestock species belong to a reduced number of linages was an unsurprising result. However, the phylogenetic overdispersion of crops was more of a shock for us, provided common knowledge on the relevance of a few agricultural plant families.

Our simple but global study should pave the way to address more directed questions. Interactions during peer-review helped us to identify additional future directions. For instance, can we use phylogenetic comparative methods to pinpoint candidates for future domestication from their phylogenetic and phenotypic profiles? , or can we ascertain if phenotypic evolution occurs at different rates or modes under farming than in the wild? In conclusion, stepping back and studying the macro-evolutionary history of domestication across taxa has and will continue to provide unique insight into the origins of our food system.