The article, titled “Finding Thermodynamically Favorable Pathways in Reaction Networks using Flows in Hypergraphs and Mixed Integer Linear Programming” and published in the Journal of Chemical Information and Modeling in 2025, introduces a mixed-integer linear programming framework that incorporates thermodynamic constraints into hypergraph-based searches for reaction pathways. In the video below, CORENET partner and co-author Prof. Daniel Merkle from the Universität of Bielefeld (UBI) discusses how the method identifies both known and novel pathways that are thermodynamically favorable and chemically plausible.

Key highlights of the paper include:

• Integrated thermodynamics into pathway search: The authors extend a hypergraph-based pathway search framework by formulating a mixed-integer linear program (MILP) that incorporates chemical potentials and molecular concentrations. This ensures that only thermodynamically favorable reaction steps are included.
• Optimization and ranking by energetics: Beyond feasibility, pathways can be ranked based on thermodynamic objective functions, like maximizing chemical work or minimizing free energy changes.
• Real-world application: The approach is demonstrated on an HCN–formamide reaction network, where it not only reproduces known literature pathways but also uncovers novel alternatives that outperform the established route under the chosen thermodynamic criteria.

Watch the video below to learn more.