Résumé
Improving root systems of fruit tree crops such as coffee is crucial to enhance water and nutrient uptake under altering temperature and rainfall patterns caused by climate change. Therefore, understanding root growth patterns and responses is key for breeding and managing coffee plants adapted to climate change and soil challenges. Emerging non-invasive imaging technologies now enable high-throughput root monitoring, opening new avenues for sustainable crop improvement. To investigate differences in coffee root system architecture and its growth dynamics and plasticity, we employed the advanced rhizotron system integrating root and shoot imaging, located in IPK's unique PhenoSphere. Eight coffee genotypes—representing two cultivated (two Arabica types and one Robusta type) and five wild species of Coffea genus—were studied under two contrasting nitrogen (N) conditions of high and low levels, with ten replicates per genotype, over a period of three months. The existing deep-learning-based image analysis pipelines were adapted for coffee roots and shoots. Principal component analysis (PCA) of the 54 extracted root traits revealed distinct clustering by species. Similarly, PCA of the 36 shoot traits showed species-specific groupings with patterns reflecting contrasting N conditions. N supply had a notable impact on RSA, particularly on traits related to root biomass: most genotypes, exhibited a significantly reduced root biovolume under low-N conditions. Nitrogen availability also influenced shoot traits; shoot dry weight revealed reductions under low-N in all species, except for one species. We are currently analyzing the time-course data to determine the timing and progression of nitrogen effects across the different coffee genotypes. In an earlier test experiment using the large automated phenotyping system with the rhizopot system, roots were sampled for exploring differences in the transcriptome and microbiome with promising first insights. A new experiment is currently conducted on the same system to generate a combined in-depth data set of HTP data, microbiome and transcriptome data that will be used for integration of all data into predictive models that will be used to select resilient genotypes with desirable root traits and nitrogen use efficiency (NUE).