A global leader in agriculture, partnered with InstaDeep to accelerate crop-trait development in corn and soybean research workflows using AgroNT.
Developing new crop traits, such as improving yield or resilience in maize, is traditionally slow and complex, often taking 10–15 years for conventional breeding and around 16.5 years for GM traits, with 5–7 years commonly spent in regulatory review. We built AgroNT to address this challenge.
AgroNT is InstaDeep’s genomic foundation model trained on genomes from 48 plant species, delivering state-of-the-art predictions for regulatory annotations, promoter and terminator strength, tissue-specific gene expression, and functional-variant prioritisation. By focusing on variants in silico, AgroNT enables teams to align experiments earlier in the pipeline and shorten discovery cycles.
Built as a DNA language model for plants, AgroNT learns the regulatory code governing gene expression and scores sequence changes for likely functional effects. In practice, AgroNT helps teams:
- Predict regulatory impact of sequence edits (e.g., promoter or terminator changes) to anticipate expression shifts before any wet-lab work.
- Prioritise variants in silico, focusing bench time on the most promising candidates and reducing the research cycle.
- Guide downstream experimentation, from construct design to follow-up assays, with sequence-level scores that support decision-making.
As our partnership progresses, insights from real-world use feed back into AgroNT’s development, strengthening model performance and deepening biological understanding over time. This collaboration shows how AI-driven genomic models can connect cutting-edge
