Interdisciplinary team awarded grant to model liquid behavior

Boise State faculty from the Department of Chemistry and Biochemistry and the Department of Computer Science were awarded a $500,000 grant from the National Science Foundation to develop computational tools for modeling liquid behavior at a molecular level. Oliviero Andreussi (chemistry) will lead the project with Konrad Meister (chemistry), Jenée Cyran (chemistry) and Edoardo Serra (computer science) as co-principal investigators. This funding will pay summer research salaries for the investigators and support two student lab assistants and one post-doctoral researcher at Boise State.

This team of researchers has come together to transform the understanding of how liquids behave at a molecular level. Existing analytical models of liquid behavior are only accurate to a point. Chemists working in fields like drug development, where understanding liquid behavior is critical when working with solutions, must improvise with analytic models based on existing empirical data or physical intuition.

Liquids like water can behave unusually in certain complex extreme conditions, like under high pressure or at extreme temperatures, and there is a significant need for more comprehensive models. Andreussi wants to use advanced techniques from computer science to bridge the gap between the current understanding and the resources modern chemists need.

The idea started with a fortuitous meeting at Boise State between Andreussi and Serra. The two began talking about Serra's research interest in graph representation learning and how it might be applied to problems in chemistry.

Graphs in this case are just collections of nodes and edges that connect them. One can imagine a social network like Facebook as a graph, with nodes representing users and edges representing people who have "friended" one another. With advanced computational techniques, researchers can perform all sorts of complex analyses on these graphs.

Andreussi and Serra will apply these techniques to chemistry, representing units of liquid as dynamic networks with nodes standing in for molecules and edges standing in for interactions between those molecules. With this approach, the researchers will create more accurate and efficient computer simulations of liquid behavior.

Although the project primarily deals in computational chemistry, it will also include lab work. Meister and Cyran both study the chemical structures associated with ice and its formation - one of those special cases that current analytic methods struggle to model. Andreussi will develop graph representation models with Serra and then test the predictions emerging from those models in the lab with Meister and Cyran.

This material is based upon work supported by the National Science Foundation under Award No. 2435964.

Share This

• Share on Facebook
• Share on X
• Share on LinkedIn
• Share through Email