UK agritech start-up Biographica, founded in 2022, is turning heads in the seed industry with an AI-driven platform designed to accelerate crop trait discovery and gene editing. Biographica claims to solve the main bottleneck in crop trait development: identifying the right genetic targets for traits like drought tolerance, disease resistance, or nutrition.
The company uses machine learning models trained on genomic and phenotype data to identify promising gene edits for traits like drought tolerance, disease resistance, and climate resilience, reducing the time and failure rates of traditional breeding pipelines.
The approach borrows directly from the pharma playbook. “We’ve seen AI reshape pharma, turning trial-and-error pipelines into learnable biological systems – and it works. We’re bringing that same discipline to crops,” said Cecily Price, CEO of Biographica.
Speed, accuracy – and targets you didn’t know existed
In pilots with leading seed and precision breeding companies, Biographica says its AI has identified proven gene targets 12× faster than traditional discovery methods with 80-100% accuracy. Beyond speed, it claims an edge in novel target discovery: finding high‑value edits legacy pipelines tend to miss.
For breeders and seed companies fighting time-to-market pressures, those metrics matter. Price estimates the platform can cut crop development timelines by up to five years and reduce R&D costs by millions – a meaningful shift for programs where every season, field trial and iteration compounds cost and risk.
Why investors and BASF | Nunhems are leaning in
Biographica has started the year strongly, raising £7 million in seed funding and announcing a new partnership with BASF | Nunhems. The capital is earmarked to expand proprietary data collection, extend the AI platform to new traits, and deepen commercial relationships across the seed industry. For strategic partners, the attraction is twofold. By moving from “trial-and-error” to learnable biological systems, Biographica pitches a predictable, scalable way to identify trait targets across crops and environments. Biographica’s platform can be fine‑tuned to partner datasets, creating what Price calls an “independent, partner‑specific discovery engine” enabling partners “to extract maximal value from their data while ensuring all data and models remain segregated”.
Scaling multi‑omics, not just sequencing
Biographica’s promise ultimately rests on data. The platform leans on proprietary multi‑omic datasets. Scaling those globally is one of the company’s biggest operational challenge. The company reduces the amount of new data needed in each domain by starting from foundation models trained on large, curated public and Biographica‑generated datasets, then fine‑tuning with targeted proprietary data.
Crucially, scaling isn’t just “more sequences,” said Price. “It’s about making datasets comparable across crops, geographies, and breeding programs.
Biographica’s team captures high‑quality sample, environment and programme metadata across crops and geographies to contextualise the data which its models are trained on. “The £7m Seed round is accelerating our ability to scale these data collection and curation pipelines,” she said.
The platform’s architecture helps reduce confidentiality concerns: by fine‑tuning models to partner data, Biographica builds an independent discovery engine for each collaboration – keeping data and models segregated. The company can also deliver projects without any partner data, removing a common barrier to engagement.
Privacy by design
A persistent tension in trait discovery is the need for large, diverse training signals versus the mandate to protect sensitive genetics. Biographica’s answer is a pre‑train, fine‑tune paradigm: train performant generalist models on diverse (non‑proprietary) datasets, then fine‑tune and deploy independently for each partner, never mixing proprietary partner data back into generalist models. Contractual controls, restricted access, and project‑bounded use of fine‑tuned models keep commercially sensitive material ring‑fenced while still letting the company scale its discovery capabilities.
Jurisdiction flexibility: discovery first, regulatory later
Given regulation remains country-specific, the platform is designed to be jurisdiction‑flexible.
Biographica sidesteps regulatory complexity by focusing on the upstream “what to change and why” discovery step, then enabling partners to deploy edits or breeding programs market by market. The platform can be customised for gene-editing or conventional breeding contexts, depending on the partner’s strategy and the regulatory environment.
Even where enabling regulation is not fully established, Price notes that many seed companies are building editing pipelines now to be ready for launch when frameworks mature. “So we can progress editing programs globally while regulations continue to evolve.”
