Skip to main content

ChemAudit

ChemAudit is a comprehensive web-based chemical structure validation suite designed for drug discovery, medicinal chemistry, and cheminformatics research. It combines powerful validation, standardization, and scoring capabilities into an intuitive interface.

What ChemAudit Does​

ChemAudit helps you ensure chemical structure quality across your research workflow:

  • Validate molecules with detailed structural checks (SMILES, InChI, MOL, or IUPAC names)
  • Screen for problematic substructures using industry-standard alert catalogs
  • Score molecules for ML-readiness, drug-likeness, and ADMET properties
  • Standardize structures using the ChEMBL-compatible pipeline with full provenance tracking
  • Process batches of up to 1 million molecules with real-time progress tracking
  • Analyze datasets with interactive charts, chemical space maps, and scaffold analysis
  • Bookmark results and track validation history with an immutable audit trail
  • Compare molecules side-by-side with property radar overlays
  • Share results via permalinks and receive notifications on completion
  • Export results in 9 formats (CSV, Excel, SDF, JSON, PDF, Fingerprints, Dedup, Scaffold, Property Matrix)

Key Features​

Single Molecule Validation​

Validate SMILES, InChI, or MOL blocks with detailed structural checks including valence, kekulization, sanitization, stereochemistry, and representation quality.

Batch Processing​

Process large datasets with up to 1 million molecules depending on deployment configuration. Real-time WebSocket progress updates keep you informed during processing.

Structural Alerts​

Screen against over 1,500 patterns from PAINS, BRENK, NIH, ZINC, and ChEMBL catalogs. Identify potential assay interference compounds and unwanted chemical moieties before investing time and resources.

Comprehensive Scoring​

Evaluate molecules across 10+ dimensions:

  • ML-readiness: Descriptor calculability, fingerprint generation, size constraints
  • Drug-likeness: Lipinski, QED, Veber, Rule of Three, Ghose, Egan, Muegge filters
  • ADMET: Synthetic accessibility, solubility, CNS penetration, bioavailability
  • NP-likeness: Natural product vs. synthetic classification
  • Scaffold analysis: Murcko scaffold extraction
  • Aggregator likelihood: Colloidal aggregation risk assessment
  • Ligand efficiency: LE and LLE metrics for lead optimization
  • Bioavailability radar: 6-axis oral bioavailability profile
  • Property breakdown: Per-atom TPSA and LogP contributions
  • Salt inventory: Salt form detection and fragment classification

ChEMBL-Compatible Standardization​

Standardize structures using a pipeline compatible with ChEMBL's curation workflow:

  • Structural issue detection
  • Salt and solvent removal
  • Parent molecule extraction
  • Optional tautomer canonicalization

QSAR-Ready Pipeline​

Prepare chemical datasets for machine learning with a multi-step curation pipeline: standardization, salt stripping, neutralization, tautomer canonicalization, and duplicate removal. Batch export in CSV, SDF, or JSON.

Structure Filter​

Multi-stage funnel filtering for generative chemistry outputs with property filters, SMARTS substructure matching, preset configurations (drug-like, lead-like, fragment-like), funnel visualization, and REINVENT-compatible scoring.

Dataset Audit​

Comprehensive dataset health auditing with an overall health score, contradictory label detection, dataset diff/comparison, curation reports, and interactive treemap drill-down.

Diagnostics​

Low-level structure analysis tools: SMILES diagnostics, InChI layer diff, round-trip validation, file pre-validation, and coordinate dimension analysis.

Batch Analytics & Visualizations​

Explore batch results with interactive charts and chemical space mapping:

  • Butina clustering with configurable Tanimoto cutoff (Morgan ECFP4 fingerprints)
  • Chemical taxonomy classification (~50 curated SMARTS rules)
  • Registration hashing for tautomer-invariant deduplication
  • Deduplication across 4 levels (exact, tautomer, stereo, salt-form)
  • Scaffold analysis with diversity metrics
  • PCA and t-SNE chemical space projections
  • Matched molecular pairs and activity cliff detection
  • Linked brushing across all visualizations

Bookmarks & History​

Save validation snapshots and maintain an audit trail:

  • Bookmark molecules with tags and notes
  • Browse validation history with filtering by date, outcome, and source
  • Submit bookmarks as batch jobs
  • Session-scoped privacy with GDPR-compliant data erasure

Scoring Profiles​

Evaluate molecules against customizable property criteria:

  • 8 built-in presets (Lipinski, Lead-like, Fragment, CNS, Ghose, Veber, PPI, NP)
  • Custom profile builder with 8 property thresholds and weights
  • Apply profiles to batch jobs or re-score subsets inline
  • Export/import profiles as JSON for sharing

IUPAC Name Support​

Enter chemical names directly — ChemAudit auto-detects IUPAC names, common names, and trade names, converting them to SMILES via OPSIN or PubChem before validation.

Notifications & Sharing​

Stay informed and share results:

  • Email notifications on batch completion
  • Webhook callbacks with HMAC-SHA256 signatures
  • Shareable permalinks for batch reports (30-day expiry)

Database Integration​

Cross-reference molecules against:

  • PubChem: Properties, synonyms, IUPAC names
  • ChEMBL: Bioactivity data, targets, clinical phase
  • COCONUT: Natural product sources and organisms
  • Wikidata: Isomeric SMILES, InChI, CAS numbers, molecular mass

Plus:

  • Identifier Resolution: Resolve 10+ identifier types (SMILES, InChI, InChIKey, CID, ChEMBL ID, CAS, DrugBank, ChEBI, UNII, Wikipedia URL, compound names) with automatic detection and cross-database linking via UniChem
  • Cross-Database Comparison: Compare structural representations across all four databases with consistency verdicts

Quick Start​

Get ChemAudit running in minutes with Docker:

# Clone the repository
git clone https://github.com/Kohulan/ChemAudit.git
cd chemaudit

# Create environment file
cp .env.example .env
# Edit .env to set required secrets (POSTGRES_PASSWORD, SECRET_KEY, etc.)

# Start all services
docker-compose up -d

Access Points (Development):

Access Points (Production):

Ready to get started? Here's where to go next:

External Resources​

Support​

Questions or issues? Open an issue on GitHub