by Tomas Veloz and Alejandro Bassi
Abstract
Identifying structural features ensuring the persistence of a reaction network is fundamental for understanding the evolution and complexification of biological entities, from origins of life to metabolic engineering. By defining organizations as subsets of species that are closed and self-maintaining, Chemical Organization Theory (COT) shows that organizations enable filtering out the regions of the phase space where attractors can exist. This result and several variants of it open a promising path to study the constitutive, developmental, and adaptive aspects of metabolic and other kinds of complex reaction networks where traditional dynamical systems methods are intractable. Despite various theoretical explorations, current methods to compute the organizations are in general limited to dozens of species/reactions due to unclear
combinatorial challenges.
In this article we study the problem of generating organizations as non-redundant combinations of smaller components. We first analyze the combinatorial challenges leading to redundant combinations. Second, we identify the minimal building components for non-redundant combinations, called Elementary Reaction Closures (ERC), and characterize them as a hierarchy. Third, we show that all non-redundant generative processes of organizations can be made aggregating ERCs that hold either a synergetic or complementarity relationship. These two notions are novel and open very interesting routes to understand the internal composition of complex metabolisms.
We next show that synergies and complementarities scale far more slowly than the structures used in previous combinatorial approaches. By computing these relationships for 438 biological reaction networks from the BioModels and BiGG databases, we obtain that ERCs grow sublinearly with reaction count while the proportion of synergies and complementarities shrinks as networks grow, confirming that larger biological networks achieve persistence through a progressively smaller set of irreducible generative relationships.
We do not prescribe algorithms to compute organizations, but discuss how algorithms based on our results could make COT applicable to reaction networks containing thousands of reactions using regular desktop computers.
Full text of the article is available at https://arxiv.org/abs/2608.03541