Skip to main content

QSAR-Ready Pipeline

The QSAR-Ready Pipeline prepares chemical datasets for machine learning by applying a standardized multi-step curation pipeline. It ensures that molecules are clean, canonical, and deduplicated before use in QSAR/QSPR modeling.

Pipeline Steps​

The pipeline applies the following steps in order:

StepDescriptionDetails
StandardizationChEMBL-compatible structure normalizationFixes nitro groups, metal disconnection, charge normalization
Salt StrippingRemove counterions and saltsExtracts the parent molecule using MolVS fragment patterns
NeutralizationNeutralize charged speciesConverts charged forms to neutral where chemically appropriate
Tautomer CanonicalizationCanonicalize tautomeric formsUses RDKit's tautomer enumerator for a canonical representation
Duplicate RemovalRemove duplicates by InChIKeyIdentifies and flags duplicate structures after standardization

Each molecule receives a result status indicating its outcome.

Result Status​

StatusMeaning
okSuccessfully curated — molecule passed all pipeline steps
rejectedFailed a pipeline step (see rejection_reason for details)
duplicateDuplicate of another molecule by standardized InChIKey
errorProcessing error (e.g., unparseable SMILES)

Using the Web Interface​

  1. Navigate to QSAR-Ready under the Data Preparation dropdown in the header
  2. Choose your input method:
    • Paste SMILES: Enter SMILES strings, one per line
    • Upload file: Drag and drop a CSV or SDF file
  3. Configure pipeline options if needed
  4. Click Process
  5. Monitor real-time progress
  6. Review results:
    • Summary statistics (ok / rejected / duplicate / error counts)
    • Per-molecule details with step-by-step provenance
    • InChIKey change tracking (original vs. standardized)
  7. Download the curated dataset in CSV, SDF, or JSON format
Batch Export

Use the download buttons to export only the successfully curated molecules (ok status) for direct use in ML pipelines.

API Reference​

Single Molecule​

Process one molecule through the pipeline:

curl -X POST http://localhost:8001/api/v1/qsar-ready/single \
  -H "Content-Type: application/json" \
  -d '{
    "smiles": "CC(=O)Oc1ccccc1C(=O)[O-].[Na+]",
    "config": {}
  }'

Response:

{
  "original_smiles": "CC(=O)Oc1ccccc1C(=O)[O-].[Na+]",
  "original_inchikey": "XAKUHBCMIFQRLG-UHFFFAOYSA-M",
  "curated_smiles": "CC(=O)Oc1ccccc1C(=O)O",
  "standardized_inchikey": "BSYNRYMUTXBXSQ-UHFFFAOYSA-N",
  "inchikey_changed": true,
  "status": "ok",
  "rejection_reason": null,
  "steps": [
    {
      "step_name": "standardization",
      "applied": true,
      "changes": ["Applied standardization rules"]
    },
    {
      "step_name": "salt_stripping",
      "applied": true,
      "changes": ["Removed fragment: [Na+]"]
    }
  ]
}

Batch Upload​

# Upload a file
curl -X POST http://localhost:8001/api/v1/qsar-ready/batch/upload \
  -F "file=@molecules.csv" \
  -F 'config={}'

# Or paste SMILES text
curl -X POST http://localhost:8001/api/v1/qsar-ready/batch/upload \
  -F "smiles_text=CCO
c1ccccc1
CC(=O)Oc1ccccc1C(=O)O"

Check Status​

curl http://localhost:8001/api/v1/qsar-ready/batch/{job_id}/status

Get Results​

curl "http://localhost:8001/api/v1/qsar-ready/batch/{job_id}/results?page=1&per_page=50"

Download Curated Dataset​

# CSV format
curl http://localhost:8001/api/v1/qsar-ready/batch/{job_id}/download/csv -o curated.csv

# SDF format
curl http://localhost:8001/api/v1/qsar-ready/batch/{job_id}/download/sdf -o curated.sdf

# JSON format
curl http://localhost:8001/api/v1/qsar-ready/batch/{job_id}/download/json -o curated.json
import requests

# Single molecule
response = requests.post(
    "http://localhost:8001/api/v1/qsar-ready/single",
    json={"smiles": "CC(=O)Oc1ccccc1C(=O)[O-].[Na+]", "config": {}}
)
result = response.json()
print(f"Status: {result['status']}, Curated: {result['curated_smiles']}")

# Batch upload
with open("molecules.csv", "rb") as f:
    response = requests.post(
        "http://localhost:8001/api/v1/qsar-ready/batch/upload",
        files={"file": f},
        data={"config": "{}"}
    )
job_id = response.json()["job_id"]

Rate Limits​

EndpointLimit
POST /qsar-ready/single30 req/min
POST /qsar-ready/batch/upload3 req/min
GET /qsar-ready/batch/*/status60 req/min
GET /qsar-ready/batch/*/results30 req/min
GET /qsar-ready/batch/*/download/*10 req/min

Use Cases​

ML Dataset Preparation​

  1. Upload your raw compound collection
  2. The pipeline standardizes, deduplicates, and cleans all structures
  3. Download the curated set and use it directly for model training

Compound Registration​

  1. Process incoming compounds through the pipeline
  2. Use the standardized InChIKey for unique registration
  3. Flag duplicates against your existing collection

Next Steps​