From computer to field: can in silico modelling can help us to optimise rice for possible droughts?

While not particularly exciting to look at, Oryza sativa (rice) is undoubtedly one of the most important plants on the planet as the major food source to 3.5 billion people. As a consequence, there is much interest in getting the most out of growing rice plants, both through maximising the yield of the plants and making them as resilient as possible to stresses they may encounter. Identifying the genetic basis of traits likely to boost rice yield or resilience to environmental stresses is complex, as these traits may be produced by a combination of different genetic loci and have low heritability. This has become easier in recent years due to advances in and wider availability of genomic and phenotyping technology. Moreover, advances in these techniques have allowed for better informing of  in silico (computer) models of plant growth and development.

Such models are of interest as they may be able to accurately measure and quantify plant responses to environmental variation through image-analysis pipelines, and to indicate how plant phenotypes respond to environmental variations. Knowledge acquired from these models could be used to actively optimise field conditions and targeted modification of rice plants to get the best out of rice in the field. In a recent study published Open Access in Journal of Experimental Botany, Malachy Campbell and colleagues model rice plant growth in relation to water availability, and use this model to identify candidate genes likely to be of interest to genetic studies of the basis of drought response in rice.

Campbell and colleagues grew different accessions (genetic variants of the same species) of rice plants and analysed their phenotypes over a 21 day period using an automated image-analysis software. Some plants were grown at 90% field capacity (100% field capacity being the point at which the soil is saturated with water) and others were not given water until they reached 20% field capacity. From the measurements they made under these high water and low water conditions, the authors produced an in silico growth model that models the relationship between shoot biomass and soil water content.

One piece of information that the authors were particularly keen to obtain from their model is the Time of Inflexion (TOI) of growth rate in response to drought. In other words, the point after which drought begins to cause reduced growth rate. Importantly, Campbell and colleagues’ model find that the TOI comes earlier for larger, faster-growing rice accessions that are experiencing drought than it does for small, slower-growing rice accessions. This supports previous work in the field showing that fast-growing accessions of rice, whilst desirable from some points of view, likely have the most to lose in times of drought.

Perhaps the most significant outcome of this study is the identification of candidate genetic regions that may influence the drought responses of rice plants. To do this, Campbell and colleagues take known genetic data about the rice accessions they use and apply it to their model. In line with observations that many growth and response traits of plants are linked to more than one genetic locus, Campbell and colleagues identified many loci with small effects on dynamic drought responses. However, the authors did consistently identify a some candidate genetic regions with larger effects on drought responses. One such region contains a gene known to be involved in responses to cold stress in rice. Whether this gene also functions in drought response is unknown, but it is worth noting that cold stress can also result in low water availability.

In identifying that faster-growing rice accessions likely suffer most quickly from drought, they highlight the importance of trying to address this in the future. As Campbell and colleagues state: ‘Further studies are necessary to determine whether these relationships can be decoupled, or to identify the optimal balance between these two attributes’. Their work also demonstrates, for the first time in rice, how in silico models can give useful indications of genetic candidates likely to merit further investigation in our understanding of plant responses to dynamic stress. Time will tell if either of these can make it to the field!