Review: Plant digital twins, the future of crop production

Imagine how cool it would be if, just like we check weather apps before going out, we could “grow” a plant in our computer before planting it into our backyard, in order to predict its future success. Pauli et al. discuss this possibility through the use of Digital Twins (DTs), virtual plants that are continuously updated with data from their real counterparts. The authors suggest that DTs could help to study crop performance and how genetic differences lead to changes in plant growth, physiology, architecture, and yield under different environments. The system could collect 3D point-cloud data (a digital representation of a 3D object), thermal and hyperspectral images, leaf traits, soil moisture, temperature, humidity, wind, and light from the real plant. An AI-enabled cyberinfrastructure system would process this data and integratae it into a functional–structural plant model. A bi-directional ray-tracing simulator can calculate the light and shading in a canopy and multiple plant DTs could be combined to make an in silico crop canopy as a crop growth model. The most ambitious idea of this paper is to predict the various biological parameters, such as flowering behavior, developmental rate, and photosynthetic capacity, use genomic predictions to create a genotype-specific virtual plant, let that virtual plant interact with different simulated environments, and observe the predicted crop performance. Overall, this paper offers a new avenue to understand genetic variation and proposes a framework that connects genes, biology, environment and crop performance is a single frame. (summary by Kavita Joshi @JoshiKvita) Trends Plant Sci. 10.1016/j.tplants.2026.06.005