A new partnership between a UK research institute and a global crop protection company is applying artificial intelligence to genomic data to tackle one of agriculture's most pressing challenges: herbicide-resistant weeds. The collaboration, funded by Innovate UK's Knowledge Transfer Partnership award, will embed genomics and AI expertise from Earlham Institute within Syngenta's research center in Berkshire to develop computational tools that predict and understand how weeds survive herbicide treatments.

Why Are Herbicide-Resistant Weeds Such a Big Problem for Farmers?

When agricultural weeds survive herbicide applications that would normally kill them, resistant populations can emerge and spread rapidly across farmland. This creates a cascading problem: farmers must use more herbicides to control the resistant plants, crops face greater competition for water and nutrients, and management becomes increasingly complex and expensive. The challenge has become critical for both individual growers and the wider agricultural sector.

One particularly difficult form of resistance is called non-target site resistance (NTSR), where weeds adapt how they absorb, move, or break down the chemical rather than developing a simple genetic mutation that blocks the herbicide's effect. Understanding and predicting this type of resistance requires analyzing complex biological patterns that traditional methods struggle to untangle.

How Will AI Genome Models Help Predict Weed Resistance?

The project will use genome large-language models (LLMs), which are large-scale AI networks trained on genome sequences to identify patterns, structures, and relationships in DNA data. Rather than requiring researchers to manually analyze genetic sequences, these AI models can process vast amounts of genomic information and predict how weeds might develop resistance mechanisms.

The team will feed existing genetic data from weed species into these AI models to help predict how weeds develop resistance to herbicides. By studying and predicting resistance patterns, researchers hope to develop methods to slow resistance spread, preserve herbicide effectiveness, and ultimately minimize excessive chemical use. This approach bridges academic expertise in genomics with real-world research and development environments where crop protection solutions are developed.

Steps to Implement AI-Driven Genomics in Agricultural Research

• Data Integration: Combine existing genetic datasets from weed species with advanced computational pipelines to create a unified resource for AI model training and validation.
• Model Development: Build and refine genome large-language models specifically designed to predict target and non-target site resistance mechanisms in agricultural weeds.
• Evidence-Based Solutions: Translate AI predictions into practical management strategies that farmers and agronomists can implement to slow resistance development and preserve herbicide efficacy.

"We have a fantastic agri-science community in the UK and this project is a perfect example of applying the latest genomics expertise to an R&D environment, helping develop evidence-based solutions for growers," said Prof Anthony Hall, Head of Plant Genomics at Earlham Institute.

Prof Anthony Hall, Head of Plant Genomics at Earlham Institute

The partnership represents a shift in how agricultural innovation happens. Rather than keeping academic genomics research separate from commercial crop protection development, the Knowledge Transfer Partnership embeds Earlham Institute's expertise directly within Syngenta's Bioscience Digital Group, which already applies data science, modeling, and computer vision to early-stage crop protection programs.

"We're delighted to be partnering with the Earlham Institute on this Knowledge Transfer Partnership. This project is an exciting opportunity to apply state-of-the-art computational approaches to tackle some of the toughest challenges in weed control," explained Dr Chris O'Grady, Senior Principal Scientist in Computational Biology at Syngenta.

Dr Chris O'Grady, Senior Principal Scientist in Computational Biology at Syngenta

The broader context for this partnership reflects a wider transformation in how AI and genomics are converging across life sciences. Illumina, a major sequencing technology company, has emphasized that genomics, multi-omics, and AI are "some of the game-changing technologies of the century," providing insights that help healthcare professionals understand disease risk, diagnose conditions earlier, and tailor treatment to individuals. The same principles apply to agriculture: understanding biological complexity at scale requires both advanced sequencing technology and AI systems that can extract meaningful patterns from massive datasets.

Meanwhile, DNAnexus, a 17-year-old platform for managing genomic and multi-omics data, has announced new AI-driven capabilities designed to make complex biological datasets more accessible and actionable. The company introduced an "Omics Data Agent," a conversational AI interface that allows researchers to create cohorts, summarize data, and run sophisticated analyses using natural language rather than requiring coding expertise. These tools reflect a broader industry trend: as genomic data accumulates faster than ever before, organizations need AI systems that can help researchers and scientists interact with that data more intuitively.

For the Earlham Institute and Syngenta partnership, the practical outcome could be significant. By developing computational pipelines and genome large-language models that predict herbicide resistance, the team aims to help farmers preserve the effectiveness of existing herbicides, reduce chemical use, and manage resistant weed populations more sustainably. The project is actively recruiting a Computational Biologist to serve as a Knowledge Transfer Partnership Associate, with a salary range of 45,000 to 48,000 British pounds and applications closing on June 12, 2026.

The convergence of genomics, AI, and agricultural science represents a shift from reactive problem-solving to predictive management. Rather than waiting for herbicide resistance to emerge and then scrambling to find solutions, researchers can now use AI models trained on genomic data to anticipate resistance mechanisms before they become widespread. This approach could reshape how the agricultural sector manages one of its most persistent challenges.