A multidisciplinary University of Liverpool-led project using artificial intelligence to identify promising molecules for applications across particle physics, cancer research and antimicrobial science has secured £8,000 funding from the Future Leaders Fellows Development Network through the Crucible Grant scheme.
Led by Dr Juri Smirnov, University of Liverpool School of Physical Science, the project brings together an interdisciplinary team including Dr Ashlea Kemp at RAL, Professor George Poulogiannis at Institute of Cancer Research and Dr Thomas Fenner at SjF Hanse Scientific GmbH to validate AI-selected molecules through laboratory work on detector materials for dark matter experiments, cancer-cell assays and antibacterial activity.
The project will investigate whether a common AI discovery engine can accelerate molecular discovery across multiple disciplines. Researchers will assess molecules identified by the platform through three linked work packages, including developing scintillator and wavelength-shifter materials for cryogenic detector systems used in dark matter experiments, screening compounds for cancer-cell inhibition and selectivity and testing antibacterial activity relevant to microbiome-related health research.
By comparing results across these distinct applications, the team hopes to understand how broadly AI-driven approaches can be applied, what generalises across disciplines, and where different scientific fields require tailored approaches.
The project will use a feedback loop in which AI suggests molecules, laboratories test them, and the results are fed back into the system to improve the next round of predictions. Through this approach, the team aims to enhance the performance of the AI platform and establish a reusable model for future AI-to-experiment research programmes.
Juri Smirnov said: “What makes this project so exciting is that we are asking whether one AI discovery platform can help solve problems across very different areas of science. Physics, cancer-cell inhibition and microbiology may seem far apart, but they all depend on finding molecules with the right properties. AI gives us a way to search that chemical space much faster and more intelligently.”
The project will focus on commercially available molecules and use staged laboratory validation, existing facilities and established partners to manage risk and maximise the chances of generating useful results.
Expected outputs include validated candidate molecules, ranked hit lists, spectra and light-yield data, a resin-based scintillator prototype, cancer-cell viability and selectivity data, antibacterial testing results, and a shared final learning report. These outputs could support future funding applications, intellectual property review, research publications and partner-led scale-up.
As well as validating the AI platform, the team aims to establish a durable interdisciplinary network linking Liverpool, national laboratories, cancer research expertise and European industry partners. Each work package is intended to seed a larger future research programme, whether in next-generation detector materials, cancer-cell inhibition or microbiome-related antimicrobial research.
The project aligns with wider efforts to explore how artificial intelligence can accelerate scientific discovery and help address challenges ranging from fundamental physics to human health. Preliminary laboratory work has already produced encouraging antimicrobial results, providing a foundation for the next phase of research.