Time‐course changes in the ionomic profiles of rice leaves and their application in growth stage prediction

Ionomic profiling of plant tissues aims to understand the role of genetic factors and external conditions in mineral nutrient composition. However, little is known about the time‐course changes occurring in these profiles during plant growth. The influences of genotype, environment, and management factors are not well understood. To clarify the variation in time‐course data and to identify factors influencing these changes, we analyzed the ionomic leaf profiles of nine rice (Oryza sativa L.) cultivars, from transplantation to harvest, under different environmental conditions. An ANOVA was conducted separately for each element to elucidate the main effects of cultivar, fertilization, and growth stage, which were highly significant for all the elements observed except fertilization. The growth stage was the most significant for all elements except B. Conversely, the fertilization effect was not significant in half of the elements studied (Li, B, Na, Mg, P, S, K, Ca, and Cd). The elements during the growth stage were relatively stable across the environments and cultivars studied. To investigate the relationship between the changing pattern and the growth stage, we predicted the growth stage of rice based on the ionomic profile leaves using a machine learning model. Over 80% of the plants in this study were correctly classified into their growth stages with 10‐fold cross‐validation using the random forest model, with a highly significant contribution of the essential macronutrients P, Ca, S, Mg, and K as explanatory variables, indicating that they could be important indicators of the growth stage of rice plants. Core Ideas The time‐course pattern of the ionome in rice leaves at the field level was revealed. The patterns of many elements were found to be affected by the growth stage. We used machine learning to uncover the relationship between ionome and growth stage. Rice growth stages could be predicted using only ionome information. The influence of macronutrients among the ionome on growth stages was shown to be significant.