Every antibody program begins with a single, high-stakes decision: which molecule should the antibody actually bind? Get it wrong and years of work and budget can follow a target that was never going to succeed. In 2025, with support from SNN under the MIT Haalbaarheid (feasibility) scheme, Formula-Y studied whether AI could make target identification faster, cheaper and more systematic.
The problem with target identification
Today, choosing a target still leans heavily on manual literature review, expert intuition and expensive, time-consuming laboratory screening. It is slow, hard to scale, and a major reason drug candidates fail later on: a great antibody against the wrong target is still a dead end.
The idea we tested
We explored an AI platform that analyses biological data to predict and prioritise the most promising, disease-relevant targets before any lab work begins. Rather than screening blindly, researchers could start from a ranked, evidence-backed shortlist. This is especially valuable in two situations: when no obvious target exists yet, as with a novel pathogen, and when many possible targets compete and the best choice is unclear.
What we examined
The study looked at the question from both sides: the technical feasibility of building and running such a platform responsibly, and the economic case, including market demand, business models, potential collaboration partners and how the resulting intellectual property could be protected.
How our approach differs
Most AI efforts in antibody discovery focus on designing or optimising the antibody itself. Far fewer address the step that comes first: systematically choosing the right target. By bringing AI to target identification as well, Formula-Y aims to complement the rest of the field rather than compete head-on with it.
Why it matters beyond the lab
Antibodies are among the most effective responses to infectious disease, autoimmune conditions and other emerging health threats. The ability to identify therapeutic targets quickly is a genuine capability for pandemic preparedness, accelerating the response in a warm outbreak phase, and building scalable groundwork in the cold phases between crises. The work aligns with the Netherlands' Health and Care innovation priorities and the Health~Holland top sector.