EVODEX: A Cheminformatics Framework for Mechanistic Abstraction and Prediction of Enzymatic Reactions
Abstract
EVODEX is a cheminformatics framework for reasoning about enzymatic chemical transformations. It represents reactions using atom-mapped structure and decomposes multicomponent enzymatic reactions into mechanistically meaningful partial transformations, remaining well-defined even when only a single reaction is known. EVODEX defines a hierarchy of abstractions based on progressively inclusive σ-bonded and π-conjugated shells around reactive centers, retaining electronic context only when it is mechanistically required. This provides a deterministic criterion for mechanistic consistency that does not rely on curated rules or statistical generalization. Applied at scale, EVODEX yields a compact set of reaction operators that support reaction validation, product enumeration, and interpretation of LC-MS data.
Fields of Study & Subjects
Machine Learning in Materials Science
Microbial Metabolic Engineering and Bioproduction
Microbial Natural Products and Biosynthesis
Publication details
- Journal / source
- Journal of Chemical Information and Modeling
- Publisher
- American Chemical Society (ACS)
- Volume
- 66
- Issue
- 8
- Pages
- 4793–4807
- Publication date
- 2 April 2026
- Language
- EN