Predictive Crop Genomics
Member of CEPLAS: Cluster of Excellence on Plant Sciences
About
We study how a plant’s genome determines its traits and develop tools to predict these traits from biological data. By applying machine learning to large-scale genomic, transcriptomic, metabolomic, and high-throughput phenotyping data, our group uncovers the molecular mechanisms driving plant performance - providing both fundamental scientific insights and new targets for genome engineering and precision breeding.
Research Topics
The central question of our research is: will a specific genotype deliver the desired trait? Answering this holds the promise of overcoming the limitations of current quantitative genetics models, though it is far from trivial. We address this challenge at several levels:
Reading and designing gene regulation. We train deep-learning models that read regulatory DNA to predict how genes are switched on and off, and how natural or engineered sequence variants change that. [click for more]
Networking multi-omics data. We develop multi-omics integration methods that link genes, transcripts and metabolites to traits like biomass, stress resilience and grain quality by predictive associations. [click for more]
Genome-metabolism interface. We use deep learning to predict how metabolites interact with proteins across the entire genome, and how genetic variance affects such interactions. [click for more]
Digital twins and self-learning plant simulations. We build digital twins of crops that capture how different genotypes adapt to changing environments, alongside self-learning simulation systems that explore how plants perform under those conditions and reveal what drives that performance. [click for more]
News
Projects
Selected Publications
Last Modified: 03.07.2026