Drug Discovery — Data Braids

Five datasets forming the core of the computational drug discovery pipeline. Together they map chemical space (PubChem), bioactivity (ChEMBL + BindingDB), screening libraries (ZINC20), and disease-specific drug sensitivity (NF Data Portal). All braided on westGate.

ChEMBL 37

2.9M compounds, 24.5M bioactivity measurements, 1.6M assays. The largest open drug discovery database mapping chemical structures to biological targets. This is the primary input for tideGlass pharmacometric modeling.

What’s Possible

• Combine with LINCS L1000 gene expression signatures for drug repurposing without wet lab access — the gen5 critical path for tideGlass
• Cross-reference with PDB for structure-based virtual screening
• Map compounds to PubChem identifiers for cross-database linking
• Feed healthSpring clinical models with bioactivity data

The Braid

sweetGrass braid.create produced a W3C PROV-O JSON-LD attestation:

{
  "@context": "https://www.w3.org/ns/prov#",
  "@id": "urn:braid:chembl-37-westgate-20260729",
  "prov:wasGeneratedBy": {
    "@type": "prov:Activity",
    "prov:used": "https://ftp.ebi.ac.uk/pub/databases/chembl/ChEMBLdb/latest/",
    "prov:wasAssociatedWith": "did:eco:westgate"
  },
  "prov:wasAttributedTo": "did:eco:westgate",
  "prov:generatedAtTime": "2026-07-29T...",
  "eco:license": "CC-BY-SA-3.0",
  "eco:blake3_root": "...",
  "eco:file_count": 2,
  "eco:size_bytes": 16106127360
}

ZINC20 SMILES (drug-like subset)

Drug-like compound subset from ZINC20 in SMILES format. Commercially available molecules filtered for drug-likeness (Lipinski rules). Virtual screening library for tideGlass.

What’s Possible

• Virtual screening library for tideGlass — compounds ready for docking
• Combine with ChEMBL bioactivity to prioritize screening candidates
• Cross-reference with PDB binding sites for structure-based screening

The Braid

sweetGrass braid.create produced a W3C PROV-O JSON-LD attestation:

{
  "@context": "https://www.w3.org/ns/prov#",
  "@id": "urn:braid:zinc20-westgate-20260729",
  "prov:wasGeneratedBy": {
    "@type": "prov:Activity",
    "prov:used": "https://zinc20.docking.org/",
    "prov:wasAssociatedWith": "did:eco:westgate"
  },
  "prov:wasAttributedTo": "did:eco:westgate",
  "prov:generatedAtTime": "2026-07-29T...",
  "eco:license": "Free for research",
  "eco:blake3_root": "...",
  "eco:file_count": 110,
  "eco:size_bytes": 167772160
}

PubChem (SMILES + InChI-Key + Synonym + Mass)

Chemical compound identifiers, structures (SMILES/InChI-Key), synonyms, and molecular masses from the world’s largest free chemistry database. The glue layer for cross-database chemical identity resolution.

What’s Possible

• Chemical identifier resolution for cross-database linking (ChEMBL, ZINC, MassBank)
• Feed MassBank spectral matching with exact masses for unknown compound identification
• Map ChEMBL bioactivity to PubChem compound metadata

The Braid

sweetGrass braid.create produced a W3C PROV-O JSON-LD attestation:

{
  "@context": "https://www.w3.org/ns/prov#",
  "@id": "urn:braid:pubchem-westgate-20260730",
  "prov:wasGeneratedBy": {
    "@type": "prov:Activity",
    "prov:used": "https://ftp.ncbi.nlm.nih.gov/pubchem/",
    "prov:wasAssociatedWith": "did:eco:westgate"
  },
  "prov:wasAttributedTo": "did:eco:westgate",
  "prov:generatedAtTime": "2026-07-30T...",
  "eco:license": "Public Domain",
  "eco:blake3_root": "...",
  "eco:file_count": 5,
  "eco:size_bytes": 11811160064
}

BindingDB binding affinity

2.9M+ binding affinity measurements (Ki, Kd, IC50, EC50) linking drug compounds to protein targets. Structure-activity relationship analysis at the binding level.

What’s Possible

• Combine with ChEMBL for comprehensive bioactivity landscape across two databases
• Cross-reference with PDB for structure-based affinity prediction
• Feed tideGlass compound ranking with experimentally measured affinities

The Braid

sweetGrass braid.create produced a W3C PROV-O JSON-LD attestation:

{
  "@context": "https://www.w3.org/ns/prov#",
  "@id": "urn:braid:bindingdb-westgate-20260802",
  "prov:wasGeneratedBy": {
    "@type": "prov:Activity",
    "prov:used": "https://www.bindingdb.org/rwd/bind/index.jsp",
    "prov:wasAssociatedWith": "did:eco:westgate"
  },
  "prov:wasAttributedTo": "did:eco:westgate",
  "prov:generatedAtTime": "2026-08-02T...",
  "eco:license": "CC-BY-3.0",
  "eco:blake3_root": "...",
  "eco:file_count": 1,
  "eco:size_bytes": 611319808
}

NF Data Portal (Synapse)

NF1 high-throughput drug screening (8K compounds + structures), NF2 Synodos drug screen, NF2 kinomics (peptide-level, protein-level, differential expression). tideGlass Module 7 — completes the 7/7 base data.

What’s Possible

• NF-specific drug repurposing: combine NF drug sensitivity with LINCS perturbation signatures for neurofibromatosis therapeutic candidates
• Cross-reference NF kinomics with ChEMBL kinase inhibitor bioactivity
• Feed tideGlass Module 7 for the NF extension of the drug repurposing pipeline

The Braid

sweetGrass braid.create produced a W3C PROV-O JSON-LD attestation:

{
  "@context": "https://www.w3.org/ns/prov#",
  "@id": "urn:braid:nf-data-portal-westgate-20260802",
  "prov:wasGeneratedBy": {
    "@type": "prov:Activity",
    "prov:used": "https://nf.synapse.org/",
    "prov:wasAssociatedWith": "did:eco:westgate"
  },
  "prov:wasAttributedTo": "did:eco:westgate",
  "prov:generatedAtTime": "2026-08-02T...",
  "eco:license": "Synapse Terms of Use",
  "eco:blake3_root": "...",
  "eco:file_count": 658,
  "eco:size_bytes": 698351616
}

See Also

• Gene Expression — LINCS L1000 (the drug repurposing partner)
• Structural Biology — PDB (binding site structures)
• What’s Possible — the drug repurposing combination