SUGAR

Systematic Utilization of Glycans for Alternate Routes: a computational platform for discovering enzymatic synthesis pathways between sugar metabolites.

What it does

SUGAR enumerates all C2-C7 monosaccharide stereoisomers, their polyol reduction products, and phosphorylated derivatives, then generates rule-based enzymatic reactions between them. A web interface lets you find optimal multi-step synthesis pathways using Yen's K-shortest-paths algorithm.

Metric	Count
Compounds	279
Reactions	1,940
Reaction types	8
Pipeline tests	116
Frontend tests	5

Quick start

Generate data

pip install -r pipeline/requirements.txt
python -m pipeline.run_pipeline

The pipeline writes compounds.json, reactions.json, and pipeline_metadata.json to pipeline/output/ and copies them to web/data/ for the frontend.

Run the frontend

cd web
npm install
npm run dev

Opens at http://localhost:3000. The pathway finder, compound browser, reaction browser, and network graph are all client-side with no backend required.

Run tests

# Pipeline (116 tests)
python -m pytest pipeline/tests/ -v

# Frontend (5 tests)
cd web && npm test

Project structure

pipeline/                        Python data pipeline
  enumerate/                     Compound generation (monosaccharides, polyols, phosphosugars)
  reactions/                     Reaction generation and cost scoring
  validate/                      Completeness, duplicate, and mass-balance checks
  import_/                       Ring 2 database enrichment (ChEBI, KEGG, RHEA, BRENDA)
  data/                          Name mappings, match overrides
  output/                        Generated JSON output
  tests/                         pytest suite

web/                             Next.js frontend
  app/                           Pages (dashboard, pathways, compounds, reactions, network, about)
  components/                    UI components (search, filters, pathway display)
  lib/                           Core logic (pathfinding, graph building, types, search)
  data/                          Copied pipeline output consumed at build time

Architecture: the Ring model

SUGAR is built in concentric rings. Each ring adds a layer of data on top of the previous one.

Ring 1 (complete) generates all compounds and reactions from first principles:

• 94 monosaccharides (63 aldoses + 31 ketoses, C2-C7)
• 41 polyols (reduction products with degeneracy detection)
• 144 phosphosugars (systematic C6 + curated biologically important compounds)
• 8 reaction types: epimerization, isomerization, reduction, phosphorylation, dephosphorylation, mutase, phospho-epimerization, phospho-isomerization
• All reactions start as "hypothetical" evidence tier

Ring 2 (implemented, optional) enriches Ring 1 data with external databases:

• ChEBI bulk matching (name, alias, formula, fuzzy)
• KEGG compound cross-referencing
• RHEA reaction matching (upgrades evidence tier to "validated" or "predicted")
• BRENDA enzyme kinetics (EC numbers, enzyme names)
• D-to-L reaction inference from validated D-form reactions

Run with database enrichment: python -m pipeline.run_pipeline (requires network access and BRENDA credentials in .env)

Run without: python -m pipeline.run_pipeline --skip-import

Ring 3 (planned) will add more derivative classes: acids, lactones, amino sugars, nucleotide sugars, deoxy sugars.

Ring 4 (planned) will add disaccharides, hypothetical enzyme engineering targets, and advanced scoring models.

Documentation

• Architecture Guide -- how the pipeline and frontend are built, how to extend them
• Data Guide -- compound/reaction data model, evidence tiers, data provenance, biochemistry rationale

Pipeline version

Current output was generated by pipeline v2.0.0 on 2026-03-20. The pipeline is deterministic: same code produces the same output.

Author

Built by the ReefBio / ReefPath Initiative team.