Database Integrations

ChemAudit integrates with major chemical databases to enrich your molecular data with cross-references, bioactivity information, natural product annotations, and structural consistency checks.

Supported Databases

PubChem

Data Available	Rate Limit
Properties, synonyms, IUPAC names, CID	30 req/min

PubChem is the world's largest collection of freely accessible chemical information. Use it to:

• Find common names and synonyms
• Get IUPAC names
• Link to PubChem compound pages
• Retrieve molecular properties

ChEMBL

Data Available	Rate Limit
Bioactivity, targets, clinical phase, ChEMBL ID	30 req/min

ChEMBL is a database of bioactive drug-like small molecules. Use it to:

• Find bioactivity data (IC50, Ki, etc.)
• Identify protein targets
• Check clinical development status
• Access published assay results

COCONUT

Data Available	Rate Limit
Natural product data, organism source, COCONUT ID	30 req/min

COCONUT (COlleCtion of Open Natural ProdUcTs) contains natural product structures. Use it to:

• Identify natural product origins
• Find organism sources
• Check natural vs. synthetic classification
• Link to COCONUT database entries

Wikidata

Data Available	Rate Limit
SMILES (isomeric), InChI, InChIKey, CAS, formula, mass	30 req/min

Wikidata is a free, open knowledge base linked to Wikipedia. Use it to:

• Get an independent structural representation from community-curated data
• Cross-check identifiers against a non-cheminformatics source
• Access CAS numbers and molecular mass
• Link to Wikidata entity pages

Isomeric SMILES

ChemAudit fetches isomeric SMILES (Wikidata property P2017) when available, which preserves stereochemistry (R/S, E/Z). If isomeric SMILES is not available, it falls back to canonical SMILES (P233).

How to Search

Web Interface

1. Enter your molecule on the Single Validation page
2. Navigate to the Database Lookup tab
3. View cross-references from all four databases:
   • PubChem properties and synonyms
   • ChEMBL bioactivity and targets
   • COCONUT natural product information
   • Wikidata identifiers and mass
4. The Cross-Database Comparison panel runs automatically (controlled by the Auto-Compare toggle), comparing structural identifiers across all databases that returned results.

Results load automatically after validation completes.

Cross-Database Comparison

After database lookup, ChemAudit compares how each database represents your molecule. This helps catch registration errors, stereochemistry inconsistencies, or identifier mismatches across public databases.

What Is Compared

The comparison checks three structural identifiers:

Identifier	Comparison Method
SMILES	RDKit canonicalization, then tautomer-invariant comparison
InChIKey	Layer-by-layer analysis (connectivity, stereo, version)
InChI	Layer analysis (formula, connectivity, hydrogen, charge, stereo)

Verdict System

Verdict	Meaning
Consistent	All compared identifiers match across databases
Minor Differences	Tautomeric or stereochemical variants of the same base structure
Major Discrepancies	Structural differences found -- different connectivity or genuinely different compounds
No Data	Fewer than 2 databases returned data, comparison not possible

The "Resolved" Entry

The comparison panel includes a Resolved entry alongside the external databases. This is ChemAudit's own RDKit-computed canonical representation of your input SMILES, providing a local reference point. It shows the kekulized SMILES form (explicit single/double bonds, no aromaticity notation).

Stereochemistry Handling

ChemAudit always prefers isomeric SMILES (preserving @/@@ stereocenters and E/Z double bonds) from each database:

• PubChem: Fetches IsomericSMILES over CanonicalSMILES/ConnectivitySMILES
• ChEMBL: canonical_smiles field includes stereochemistry by default
• COCONUT: canonical_smiles field includes stereochemistry by default
• Wikidata: Fetches P2017 (isomeric SMILES) over P233 (canonical SMILES)

This ensures that stereochemical information is preserved in all comparisons.

API - Cross-Database Comparison

curl -X POST http://localhost:8001/api/v1/integrations/compare \
  -H "Content-Type: application/json" \
  -d '{"smiles": "C[C@H](N)C(=O)O"}'

Response:

{
  "entries": [
    {
      "database": "PubChem",
      "found": true,
      "canonical_smiles": "C[C@@H](C(=O)O)N",
      "inchikey": "QNAYBMKLOCPYGJ-REOHCLBHSA-N",
      "inchi": "InChI=1S/C3H7NO2/..."
    },
    { "database": "ChEMBL", "found": true, "..." : "..." },
    { "database": "COCONUT", "found": true, "..." : "..." },
    { "database": "Wikidata", "found": true, "..." : "..." },
    { "database": "Resolved", "found": true, "..." : "..." }
  ],
  "comparisons": [
    {
      "property_name": "canonical_smiles",
      "status": "match",
      "detail": "Raw SMILES differ but resolve to same canonical form after RDKit canonicalization."
    }
  ],
  "overall_verdict": "consistent",
  "summary": "All compared properties match across databases"
}

Identifier Resolution

ChemAudit can resolve a wide range of chemical identifiers to a canonical SMILES structure.

Supported Identifier Types

Type	Example
SMILES	CCO, C[C@H](N)C(=O)O
InChI	InChI=1S/C2H6O/c1-2-3/h3H,2H2,1H3
InChIKey	LFQSCWFLJHTTHZ-UHFFFAOYSA-N
PubChem CID	702
ChEMBL ID	CHEMBL545
CAS Number	64-17-5
DrugBank ID	DB00898
ChEBI ID	CHEBI:16236
UNII	3K9958V90M
Wikipedia URL	https://en.wikipedia.org/wiki/Ethanol
Compound name	aspirin, caffeine

API - Resolve Identifier

curl -X POST http://localhost:8001/api/v1/resolve \
  -H "Content-Type: application/json" \
  -d '{"identifier": "CHEMBL25"}'

Response:

{
  "resolved": true,
  "identifier_type_detected": "chembl_id",
  "canonical_smiles": "CC(=O)Oc1ccccc1C(=O)O",
  "inchi": "InChI=1S/C9H8O4/...",
  "inchikey": "BSYNRYMUTXBXSQ-UHFFFAOYSA-N",
  "iupac_name": "aspirin",
  "resolution_source": "chembl",
  "resolution_chain": ["chembl_id → ChEMBL API → SMILES"],
  "cross_references": {
    "pubchem_cid": 2244,
    "chembl_id": "CHEMBL25"
  },
  "confidence": "high"
}

API - PubChem Lookup

curl -X POST http://localhost:8001/api/v1/integrations/pubchem/lookup \
  -H "Content-Type: application/json" \
  -d '{"smiles": "CCO"}'

Response:

{
  "found": true,
  "cid": 702,
  "iupac_name": "ethanol",
  "molecular_formula": "C2H6O",
  "molecular_weight": 46.07,
  "synonyms": ["ethanol", "ethyl alcohol", "alcohol"],
  "url": "https://pubchem.ncbi.nlm.nih.gov/compound/702"
}

API - ChEMBL Bioactivity

curl -X POST http://localhost:8001/api/v1/integrations/chembl/bioactivity \
  -H "Content-Type: application/json" \
  -d '{"smiles": "CCO"}'

Response:

{
  "found": true,
  "chembl_id": "CHEMBL545",
  "pref_name": "ETHANOL",
  "max_phase": 4,
  "bioactivity_count": 1250,
  "bioactivities": [
    {
      "target_chembl_id": "CHEMBL240",
      "target_name": "GABA receptor",
      "activity_type": "IC50",
      "activity_value": 100.0,
      "activity_unit": "nM"
    }
  ],
  "url": "https://www.ebi.ac.uk/chembl/compound_report_card/CHEMBL545"
}

API - COCONUT Lookup

curl -X POST http://localhost:8001/api/v1/integrations/coconut/lookup \
  -H "Content-Type: application/json" \
  -d '{"smiles": "CCO"}'

Response:

{
  "found": true,
  "coconut_id": "CNP0123456",
  "name": "Compound Name",
  "smiles": "CCO",
  "molecular_weight": 46.07,
  "organism": "Genus species",
  "url": "https://coconut.naturalproducts.net/compounds/CNP0123456"
}

Understanding Results

PubChem Results

CID (Compound ID): Unique PubChem identifier

IUPAC Name: Systematic chemical name

Synonyms: Common names, trade names, registry numbers

URL: Direct link to PubChem compound page

Synonym Search

PubChem synonyms help identify whether a molecule is a known drug, reagent, or natural product.

ChEMBL Results

ChEMBL ID: Unique ChEMBL identifier

Preferred Name: ChEMBL's canonical name for the compound

Max Phase: Clinical development status:

• 0: Preclinical
• 1: Phase I clinical trials
• 2: Phase II clinical trials
• 3: Phase III clinical trials
• 4: Approved drug

Bioactivity Count: Number of bioactivity records

Bioactivities: Sample bioactivity data (up to 10 results):

• Target ID and name
• Activity type (IC50, Ki, EC50, etc.)
• Activity value and unit

Clinical Phase

Max phase 4 indicates an approved drug. Phase 0 means the compound has bioactivity data but is not in clinical development.

COCONUT Results

COCONUT ID: Unique identifier in the COCONUT database

Name: Natural product name

Organism: Source organism (plant, fungi, bacteria, etc.)

URL: Direct link to COCONUT database entry

Rate Limits

All database integrations are rate-limited to prevent abuse of external services:

Database	Anonymous Limit	Authenticated Limit
PubChem	30 req/min	30 req/min
ChEMBL	30 req/min	30 req/min
COCONUT	30 req/min	30 req/min
Wikidata	30 req/min	30 req/min

Rate Limiting

If you exceed rate limits, you'll receive a 429 error. Wait before retrying, or use batch processing which handles rate limiting automatically.

Handling Missing Data

Not all molecules are in all databases. The API returns found: false when no match is found:

{
  "found": false,
  "message": "No match found in PubChem"
}

This is normal for:

• Novel compounds
• Synthetic intermediates
• Proprietary structures
• Very rare natural products

Use Cases

Compound Identification

Lookup molecules to confirm identity:

1. Validate structure
2. Cross-reference against databases
3. Compare synonyms and IDs with your records
4. Verify molecular formula and weight

Bioactivity Screening

Check if molecules have known bioactivity:

1. Search ChEMBL
2. Review targets and activity types
3. Check clinical development status
4. Access published data for context

Natural Product Analysis

Identify natural product origins:

1. Search COCONUT
2. Check organism source
3. Review natural product classification
4. Link to literature references

Integration in Workflows

Batch Processing

Database lookups can slow batch processing. Use them selectively:

# Process batch without database lookups (faster)
curl -X POST http://localhost:8001/api/v1/batch/upload \
  -F "file=@molecules.sdf"

# Then lookup specific molecules of interest individually

Compound Curation

Enrich your database with cross-references:

1. Validate and standardize structures
2. Lookup each database
3. Store IDs (CID, ChEMBL ID, COCONUT ID)
4. Link to external pages for detailed information

Best Practices

1. Cache results: Database APIs are rate-limited, cache lookups to avoid repeated requests
2. Handle failures gracefully: External APIs can be temporarily unavailable
3. Verify matches: Always verify the returned structure matches your input
4. Respect rate limits: Implement exponential backoff for retries
5. Use InChIKey for lookups: More reliable than SMILES for exact matching

Next Steps

• IUPAC Name Conversion — Enter chemical names directly instead of SMILES
• Exporting Results — Export data with database cross-references
• Batch Processing — Process large datasets
• API Reference — Full integrations API documentation