About SUGAR
SUGAR (Sugar Utilization and General Anabolic Routes) is a computational database and pathway explorer for carbohydrate metabolites and their enzymatic interconversions. It enumerates monosaccharides, polyols, phosphorylated sugars, and related compounds from first principles of stereochemistry, then generates a reaction network connecting them via known and predicted enzymatic transformations.
The goal of SUGAR is to make the full landscape of sugar biochemistry navigable: from finding synthetic routes between rare monosaccharides, to understanding which reactions are well-supported by experimental evidence versus which are computationally inferred. Every compound, reaction, and pathway is tagged with provenance metadata so researchers can calibrate their trust in each result.
The web interface provides a compound browser, reaction browser, k-shortest-path finder, and an interactive network graph — all filterable by evidence tier so you can focus on only the reactions you trust.
Methodology
Compound enumeration
Monosaccharides are enumerated by systematically assigning R/S stereocenters to a carbon backbone of length 2–7 while enforcing chemical valence rules. The enumeration covers all aldoses, ketoses, deoxy sugars, amino sugars, and their corresponding polyols and lactones. Each compound is assigned a canonical identifier, molecular formula, and stereocenters list. A separate pass resolves common names and aliases from a curated synonym table (covering IUPAC names, trivial names, and CAS synonyms).
Reaction generation
Reactions are generated by applying reaction templates to compatible substrate pairs. Templates encode the bond-breaking and bond-forming logic for each reaction class (epimerization, isomerization, reduction/oxidation, phosphorylation, nucleotidyl-transfer, etc.). Each template produces a candidate reaction which is then checked for thermodynamic plausibility using estimated ΔG values. Reactions matching entries in RHEA, BRENDA, or UniProt are promoted from hypothetical to validated or predicted tiers based on the quality of the match.
Pathway finding
The reaction network is stored as a directed graph where nodes are compounds and edges are reactions weighted by a cost score (lower is better). Pathway finding uses Yen's k-shortest paths algorithm to enumerate diverse routes between a source and target compound. Results can be filtered by evidence tier so that, for example, only pathways composed entirely of validated reactions are shown.
Evidence Tiers
Each reaction is assigned an evidence tier reflecting the strength of experimental or computational support.
Experimentally confirmed reaction with published kinetic or structural data (EC number, PMID, or BRENDA entry).
Computationally predicted from homologous enzyme sequences or known reaction families with high confidence.
Inferred from pathway context, thermodynamic feasibility, or partial experimental evidence.
Chemically plausible reaction enumerated from stereochemical rules with no direct experimental support.
Data Version
- Pipeline version
- 2.0.0
- Generated
- March 26, 2026
- Total compounds
- 316
- Total reactions
- 2,096
- Monosaccharides
- 94
- Polyols
- 41
- Epimerizations
- 478
- Isomerizations
- 124
- Reductions
- 94