Domesticated species are excellent model systems to understand different types of evolutionary forces, a point not lost on Darwin when he launched his pioneering work. Over the last few centuries, attention has focused on understanding how domesticated species arose from their wild ancestors; in contrast, relatively little attention has been paid to how crops species dispersed from their centers of origin into new habitats as they spread across the globe. 

Our group has studied different aspects of domesticated crop species evolution and genetics for nearly 20 years. As genomic data became abundant, we saw the opportunity to reconstruct the genomic history of crop dispersal by combining genomic and geographic information. We recently reported on just such a reconstruction for the dispersal of Asian rice, which we published recently in Nature Plants.

 Rice (Oryza sativa) is one of the key crop species of the world, a staple for about half of the world’s population. This species was domesticated starting about 9,000 years ago in the Yangtze Valley in China, and from there spread across much of East, Southeast and South Asia, and later into the Middle East, Africa, Europe and the Americas. In the process of this dispersal, rice populations had to adapt to different environments and cultural milieus, and it is this process of adaptation that intrigued us.

 One of us (Rafal Gutaker) had used genomics to reconstruct the dispersal of multiple crop plants, so when he joined the Purugganan laboratory as a postdoc he was given the challenge of reconstructing the dispersal of rice. He was able to use data from the large Rice 3K Genome dataset, as well as whole genome re-sequencing data that had been sequenced at the Purugganan lab in collaboration with Ines Pires, Margarida Oliveira and Sonia Negrao at ITQB Lisbon. Using admixture graph approaches on genome sequence data from more than 1,400 rice landraces, Rafal was able to reconstruct the relationships between rice populations found in different geographic areas, as well as use coalescent methods to date the times when populations split.

Seeking to expand his approach, Rafal wanted not only to examine genome data, but also explore rice ecology using available environmental and geographic data. We needed an expert in that area, so we reached out to Jesse Lasky at Penn State and began to collaborate on this idea. Jesse’s work yielded unexpected results. We all thought that water availability (precipitation) would be the most limiting factor in rice variation distribution, but it turned out to be temperature instead. More precise analyses revealed that gradients in heat accumulation and temperature were very strongly associated with the genomic differences between tropical and temperate japonica rice varieties.

 At about the same time, we became aware of the work of Jade d’Alpoim Guedes from the University of California at San Diego. Jade’s work has been to combine archaeology with paleoclimate modeling in China to examine the effects of climates on ancient agriculture. From their work, Jade and her collaborator Kyle Bocinsky had proposed that two cooling events about 5,000 and 4,200 years ago forced tropical rice movement from it’s domestication center and led to the emergence of temperate varieties.

 It was very exciting when we compared their analysis with our estimation of past demography from genomic data. We observed a severe bottleneck in tropical rice populations at around 4,200 years BP, at the moment when Jade and Kyle’s analysis showed a paleoclimate cooling event.

 One piece of puzzle that we were missing, however, was archaeological data. For those of us who study crop evolution, archaeology is our palaeontology – a way to examine fossil remains and what they tell us about evolution.

 Over the last decade, the Purugganan lab had worked with the archaeobotanist Dorian Fuller from University College London on multiple projects that weaved genetics and archaeology in the study of crop domestication. We reached out to Dorian who was excited to hear about our project and shared with us the comprehensive database of archaeological rice remains excavated in Asia starting from about 9,000 years ago. When we compared the geographic distribution of rice before and after the cooling event 4,200 years ago, we realised it matched expectations from the paleoenvironmental model - tropical rice migrated southwards, while local rice slowly adapted to northern latitudes as temperate varieties.

 It all came together – the genomic sequence data from 1,400+ rice landraces gave us a model for where and when rice spread to different parts of Asia, the archaeology gave us “ground truth” information on when and where rice shows up at various places, and the environmental and paleoclimate modeling gave us the ecological context associated with rice dispersal. Together, this multidisciplinary approach allowed us to write a first draft of the story of how rice dispersed across Asia.

Figure 1. Pattern and timing of the dispersal of japonica rice, based on genomics, archaeobotany and paleoclimate modeling.

 Understanding the environmental pressures over the last 10,000 years of climate fluctuations, and colonization of new habitats by domesticated species, could help us in developing new crop varieties that meet new environmental challenges of the future, helping address looming food security issues.