About EVODEX.1
EVODEX is a framework for understanding how enzymes transform molecules. It helps researchers figure out which products can be formed from a given starting molecule, determining the mechanistic plausibility of potential transformations, and interpreting biochemical data. This is useful in areas like metabolic engineering, drug discovery, interpreting lab data, and enabling AI to reason about biochemistry.
EVODEX.1 is a release of the EVODEX (Enzymatic Validation Operator Differentiation and Extraction) framework that abstracts enzymatic reactions derived from the BRENDA database to generate a family of named operators. These include representations of full and partial reactions, mass and formula changes, and mechanistic patterns derived from sigma and pi bonding interactions. Additionally, it provides algorithms for using the operators. For information on using EVODEX, visit our GitHub repository.
The “.1” indicates that this is version 1 of EVODEX. All operator identifiers are namespaced with this version number (e.g., EVODEX.1-R12), so they can be uniquely referenced and distinguished from past and future versions.
For more technical information on EVODEX, see our preprint at https://www.biorxiv.org/content/10.1101/2025.06.19.660615v1.full.
Google Colab Notebooks
These interactive notebooks demonstrate how EVODEX.1 operators can be applied to predict biosynthetic pathways, evaluate proposed reactions, and interpret mass spectrometry data. Each demo walks through real examples using operators mined from the EVODEX.1 dataset.
EVODEX.1 Operator Glossary
| Operator | Description |
|---|---|
| R | Full atom-mapped reaction |
| P | Partial reaction based on shared atom maps |
| F | Formula difference between reactants/products |
| M | Mass difference for MS analysis |
| C/N/E | Core, Nearest-Neighbor, and Electronic operators |
| Cm/Nm/Em | Matched versions using only mapped atoms |
All Operators by Type
These links provide a browsable, website-based version of the EVODEX.1 dataset. Each page is internally hyperlinked, allowing you to explore relationships between operators and navigate between related abstractions. If you prefer working offline or programmatically, the full set of operators is also available for download as CSV files in our GitHub data directory. When using EVODEX via PyPI, these operator files are included with the library.
Focused Subsets for Synthesis and Mass Spectrometry
These focused subsets of the EVODEX.1 dataset support specific EVODEX use cases. The synthesis subset includes EVODEX-E operators filtered to remove ubiquitous cofactors simplifying the process of predicting biosynthetic pathways in a cellular context. The mass spectrometry subset is derived from EVODEX-M and includes only transformations consistent with mass shifts observed between single ions, that is, where either the substrate or product consists of a single molecular species. This restricts the set to transformations that would plausibly relate the masses of two peaks in an LCMS trace through a substrate/product relationship, enabling mechanistic interpretation of observed mass differences.
Credits and Contact
Developed by the Anderson Lab, Department of Bioengineering, UC Berkeley.
For more information, contact: jcanderson@berkeley.edu
Please cite our work when using EVODEX. Citation info and DOI will be added here once the associated paper is published.
EVODEX was derived from the following resources:
- BRENDA: The Comprehensive Enzyme Information System
- EnzymeMap: Heid E, Probst D, Green WH, Madsen GKH. EnzymeMap: curation, validation and data-driven prediction of enzymatic reactions. Chem Sci. 2023;14(48):14229–14242. PMID: 38098707; PMCID: PMC10718068.