In Search of Chemical Reaction Network Science.
Evan Spotte-Smith, University College Dublin.
29th of January 2026
Abstract:
Chemical reaction networks (CRNs) — collections of interrelated species and interdependent reactions — are ubiquitous and powerful objects of study, even if many chemists shy away from the complexity that CRNs represent. In this presentation, I will introduce CRNs and discuss my work computationally modeling these chemo-mathematical structures. Following initial definitions of CRNs and a brief description of their applications, I will focus on different attempts to represent CRNs, from differential equations to simple graphs and bipartite graphs. With this context, I will argue for the possibility and necessity of representing CRNs as directed hypergraphs, generalizations of directed graphs that allow for beyond-binary relations (directed hyperedges) between entities (vertices). With the remainder of my time, I will discuss my (very) preliminary work at the intersection of hypergraph theory and CRNs, including developing a software framework in Julia for directed hypergraphs (SimpleDirectedHypergraphs.jl); designing a random model for CRN-like directed hypergraphs; and working towards machine learning on CRN hypergraphs structures. I encourage critical discussion, including possibilities of collaboration and critique of these nascent projects.