Skip to main content
Go to the home page of the European Commission (opens in new window)
English en
CORDIS - EU research results
CORDIS

AI-powered classification of bimolecular reaction mechanisms from kinetic data

Project description

Smart tools for sustainable bimolecular chemistry

Understanding how molecules react is vital for developing green technology, but traditional methods are slow and often resource-intensive. Supported by the Marie Skłodowska-Curie Actions programme, the MecAI project aims to address this by developing an easy-to-use AI tool that predicts complex bimolecular reaction mechanisms directly from experimental data. Building on high-accuracy preliminary research, the project will extend this framework to cover more than 1 000 distinct chemical pathways, including various activation and deactivation steps. Designed for accessibility, the tool aims to support non-technical users in their investigations, surpassing current state-of-the-art capabilities. Ultimately, this innovation could replace traditional hit-or-miss screening, accelerating chemical discovery and fostering more efficient, sustainable chemistry.

Objective

Mechanistic studies in chemistry are fundamental to deepen our knowledge of how organic molecules behave and react, facilitate the optimisation of chemical processes, design innovative catalysts, explore new reactivity, and progress toward greener and more efficient chemistry, while avoiding the risks of traditional hit-or-miss screening methods. Yet, such studies typically demand substantial time and resources and can still yield inconclusive results. These limitations create an urgent need for more efficient approaches, and the rapid rise of artificial intelligence tools developed across many fields is opening powerful new possibilities to meet that need. Proof-of-concept work from Larrosa and Burés shows the potential of AI-based models as powerful tools for assisting kinetic analyses and elucidating reaction mechanisms. In their 2023 Nature paper, they show the possibility of training a deep learning model with data containing a variety of kinetic profiles from several mechanisms and demonstrated up to 99.99% accuracy, even under conditions with simulated experimental errors. While this proof-of-concept work only focuses on a small set of 20 unimolecular reaction mechanisms, our goal is to develop a simple-to-use tool powered by artificial intelligence that predicts bimolecular reaction mechanisms from experimental kinetic data. This project builds on the strong basis established in the published preliminary work and seeks to extend the framework to a new model capable of handling bimolecular reactions and encompassing over 1,000 distinct mechanisms with various activation and deactivation pathways. This tool will be designed for accessibility and will support non-technical users in their mechanistic investigations, and will dramatically surpass current state-of-the-art capabilities.

Fields of science (EuroSciVoc)

Keywords

Programme(s)

Topic(s)

Funding Scheme

HORIZON-TMA-MSCA-PF-EF - HORIZON TMA MSCA Postdoctoral Fellowships - European Fellowships

See all projects funded under this funding scheme

Call for proposal

(opens in new window) HORIZON-MSCA-2025-PF

See all projects funded under this call

Coordinator

THE UNIVERSITY OF MANCHESTER
Net EU contribution
€ 260 347,92
Address
OXFORD ROAD
M13 9PL Manchester
United Kingdom
Region
North West (England) Greater Manchester Manchester
Activity type
Higher or Secondary Education Establishments
Links
Total cost
No data