In recent years, there has been a growing interest in the stress-tolerant oilseed crop Camelina sativa. However, despite this interest, there has been relatively little research into the plasticity of this species, which has not been subject to significant breeding efforts. To address this knowledge gap, the EU H2020 UNTWIST project (https://www.untwist.eu/) is exploring the diversity of stress response mechanisms in different lines of camelina and using this new understanding to develop predictive models.In this study, we conducted targeted and untargeted metabolomics analyses on 54 camelina lines grown under three different conditions (control, heat, and drought stress). We quantified ten major metabolic traits and used an untargeted LC-MS metabolomics approach to characterise 3,016 metabolic features. The same 54 lines were also evaluated (phenology, physiology and performance parameters) in field trials across Europe (France, England, Italy and Spain). Our findings show that the different camelina lines exhibited a wide range of stress responses, which were not always linked to their genetic background. Using predictive metabolomics, we were able to accurately predict important agronomic and physiological variables, including thousand-kernel-weight and fatty-acid composition, based on data obtained from plants grown in the greenhouse. Our results suggest that metabolites present in early-stage leaves of camelina plants contain information about future performance of the lines. Annotation of the metabolic biomarkers identified in this study will give insights into the biology behind those predictions.< Réduire