{
  "source": "ORD Dataset Wishlist",
  "snapshot_date": "2026-09-25",
  "ord_wishlist_status_at_snapshot": "Available",
  "ord_wishlist_card": "https://github.com/orgs/open-reaction-database/projects/11?pane=issue&itemId=126570023",
  "ord_data_issue": null,
  "title": "Efficient large-scale exploration of fragment hit progression by exploiting binding-site purification of actives (B-SPA) through combining multi-step array synthesis and HT crystallography.",
  "doi": "10.1002/anie.202424373",
  "journal": "ACIE",
  "publication_date": "2025-02-11",
  "authors": [
    "Harold Grosjean",
    "Anthony Aimon",
    "Storm Hassell‐Hart",
    "Warren Athol Thompson",
    "L. Koekemoer",
    "James M. Bennett",
    "Anthony R Bradley",
    "C.A. Anderson",
    "Conor Francis Wild",
    "William J. Bradshaw",
    "Edward A. Fitzgerald",
    "Tobias J. Krojer"
  ],
  "article_open_access": {
    "status": "hybrid",
    "license": "cc-by"
  },
  "dataset_size_estimate": 1876,
  "experiment_type": "HTE synthesis",
  "ord_priority": "Medium",
  "ord_difficulty": "Normal",
  "source_data_kind": "open_repository_machine_readable",
  "smiles_in_source": "yes",
  "needs_author_contact": false,
  "data_locations": [
    {
      "url": "https://zenodo.org/records/7586212",
      "what": "Zenodo record (concept DOI 10.5281/zenodo.7586211; record title uses the earlier manuscript title 'High-throughput crystallography for rapid early-stage fragment growth from crude arrays by low-cost robotics'). Files: 'X-ray and reaction QC data summary.csv' (1.0 MB, 2262 rows; columns include iteration/series, Smiles final product, Reactant 1-3 name+SMILES, Synthesis/LCMS outcome, Yield, QC comment, plus crystallography/biophysics columns), 'Opentrons scripts - protocols used for high throughput synthesis.zip' (1.7 MB), 'MSCheck - semi-automated QC analyser results and reports.zip' (32 MB).",
      "license": "CC-BY-4.0"
    },
    {
      "url": "https://doi.org/10.26434/chemrxiv-2023-6m2s0-v2",
      "what": "ChemRxiv preprint (procedures/conditions)",
      "license": "unknown (not checked)"
    },
    {
      "url": "https://doi.org/10.1002/anie.202424373",
      "what": "Angewandte Chemie article (Wiley; not fetched)",
      "license": "CC-BY (article, per OpenAlex)"
    }
  ],
  "notes": "This is the B-SPA fragment-progression study (PHIP(2)). Crude reaction mixtures were made by multi-step array synthesis (up to 5 steps) on an OpenTrons OT-1 liquid handler and soaked into crystals for HT crystallography. Source data: Zenodo record 7586212 (CC-BY-4.0). 'X-ray and reaction QC data summary.csv' has 2262 rows, and every row has product SMILES. Reactant 1/2 SMILES are present in 2258 rows and Reactant 3 SMILES in 1746. The Synthesis/LCMS outcome column is qualitative. Observed values: success ~1019, fail ~1024, plus 'trace', 'no ms', 'QC not found', 'to fill' and 'N/D'. The Yield column is mostly '0%', 'not calculated' or blank. Conditions are not in the CSV. Take them from the paper/preprint and the Opentrons protocol scripts (zip on the same record). Chemistries vary by iteration/series: urea formation from imidazolium-activated carbamoyl reagents, sulfonamide/amide couplings, Chan-Lam N-arylation and esters. Expect one reaction template per series. Crystallography, PanDDA and biophysics columns (Kd etc.) are not reaction data and should not be encoded as reaction outcomes. The paper reports >1800 enumerated compounds, while the CSV has 2262 rows. Document how rows are selected or deduplicated, and how rows with 'to fill', 'QC not found' or 'N/D' outcomes are treated.",
  "verification_caveats": "Outcomes are qualitative (LCMS product detected yes/no/trace), with no yields for most rows. Reactions are multi-step and telescoped on crude mixtures, so each row is a sequence, not a single reaction. There are several reaction templates across iterations. Some reagents are salts (imidazolium iodides). Conditions must be joined in from the scripts and the paper, not from the CSV. Some outcome cells are placeholders ('to fill')."
}