THE LATEST
British researchers develop a machine-learning model to help discover new cryoprotectants for cold storage
The research conducted by scientists from the UK has led to the development of a cutting-edge computational framework that enhances the safe freezing of essential medicines and vaccines. This innovative approach, outlined in the academic journal Nature Communications, marks a significant advancement in cryopreservation, improving the viability and effectiveness of crucial healthcare treatments.
Cryopreservation is essential for storing vaccines, fertility materials, blood donations, and other therapies. It relies on specialized molecules called "cryoprotectants" to maintain the integrity and stability of stored materials during freezing, preserving their therapeutic properties. Without effective cryopreservation methods, treatments may need to be used immediately, limiting their availability for future use.
The research team, led by Professor Gabriele Sosso of the University of Warwick, used machine learning to test hundreds of new molecules as potential cryoprotectants. According to Prof. Sosso, the model's success came from its collaboration with traditional methodologies, demonstrating the value of integrating machine learning with molecular simulations and experimental work.
A significant finding was the identification of a novel molecule capable of inhibiting ice crystal growth during freezing, addressing a longstanding challenge in cryopreservation. While existing cryoprotectants protect cells, they often fail to prevent ice crystal formation, which can compromise the integrity of stored materials.
Dr. Matt Warren, a PhD student involved in the project, highlighted the transformative impact of the machine learning model in predicting cryoprotective activity. He emphasized the potential of machine learning to accelerate scientific research and streamline data analysis.
The team's experiments, including the demonstration of reduced cryoprotectant volumes needed for blood storage and transfusion processes, highlight the practical implications of their findings. These results not only promise to expedite the discovery of new cryoprotectants but also potentially repurpose existing molecules to enhance ice growth inhibition.
Professor Matthew Gibson of the University of Manchester praised the collaboration with Prof. Sosso and emphasized the groundbreaking nature of the findings. He noted that the computational model's identification of active molecules represents a significant leap in understanding cryoprotective properties, showcasing the transformative potential of machine learning in scientific discovery.
This study opens the door to accelerated advancements in cryopreservation research, offering new avenues for the development of efficient cryoprotectants with far-reaching implications across various healthcare sectors.
