Nextmol is a computational chemistry dedicated to the development of advanced modelling and AI tools for the chemical industry. A key focus for the company is to enable faster development of sustainable chemicals. In 2024, Nextmol was awarded 30,000 node hours on EuroHPC supercomputer MareNostrum 5 to employ Machine Learning (ML) to screen candidate biopolymers. Biopolymers (polymers developed from natural sources) have emerged as a promising solution to replacing petrochemical polymers and reducing plastic pollution.
Current research suggests that blending biopolymers will lead to a more durable product with similar advantages to plastic. Testing thousands of possible biopolymer combinations through experimentation in an experimental lab is slow and costly. High-performance computing provides an excellent route to finding a solution more quickly.
With the ability to perform 314 billion calculations per second, MareNostrum 5 has the compute power to allow the research scientists from Nextmol rapidly create essential datasets. Nextmol performed high-throughput Molecular Dynamics simulations of over 500 polymers and polymer blends to generate an extended and consistent computational dataset of biopolymer properties that has subsequently been used for the training of a predictive Machine Learning model.
Nextmol will continue to build on the discoveries from this project and the data and Machine Learning model will be utilised to further determine structure-property relationships of biopolymers and for analyses of synergistic/antagonistic effects. Data produced from this research will be used to create new biopolymers with the potential to deliver sustainable alternatives to petroleum-based polymers.
Nextmol will present this work at the EuroHPC User Days 2025 on 30/09-01/10/2025 in Copenhagen, Denmark, during the Chemical Sciences parallel session on Day 1 of the event.
