Optimizing Model Learning Performance on a Challenging Heck Reaction Yield Data Set

{
  "problem": {
    "id": "prb_uimaq8zt7i",
    "slug": "ord-wang-2025-hecklit-heck-yields",
    "title": "Optimizing Model Learning Performance on a Challenging Heck Reaction Yield Data Set",
    "structure": "SOLVE",
    "status": "open",
    "campaign": "open-reaction-database-wishlist",
    "summary": "Curate this paper's reaction data into a validated Open Reaction Database dataset. JOC 2025, DOI 10.1021/acs.joc.5c01641. About 50,000 reactions · literature review. ORD wishlist priority High, difficulty Easy. Source data: machine-readable files in an open repository."
  },
  "revision": {
    "id": "prv_8yyztp5sfp",
    "revision_number": 1,
    "created_at": "2026-09-25T19:27:21Z",
    "changelog_md": null,
    "statement_md": "# Curate this paper into an ORD dataset\n\nBuild an Open Reaction Database (ORD) dataset of the reactions reported in:\n\n- **Paper:** Optimizing Model Learning Performance on a Challenging Heck Reaction Yield Data Set\n- **Authors:** Shen Wang, Yining Liu, Weiren Zhao et al.\n- **Published:** JOC, 2025-09-03\n- **DOI:** [10.1021/acs.joc.5c01641](https://doi.org/10.1021/acs.joc.5c01641)\n- **ORD wishlist card:** [open on GitHub](https://github.com/orgs/open-reaction-database/projects/11?pane=issue&itemId=128982646)\n\nThe reference input `card` holds these facts, the source-data locations and the notes below in machine-readable form.\n\n## Scope\n\nEvery reaction the paper reports with its data (the wishlist estimates about 50,000 reactions; literature review): inputs with amounts and roles, conditions, workups, outcomes and provenance, following the `ord-curation` Skill. Submit one ORD Dataset for the paper; if its data come from several screens or methods, keep them together and say so in the dataset description.\n\n## Source data\n\nSource: J. Org. Chem. 2025, DOI 10.1021/acs.joc.5c01641 (HeckLit, a literature-derived Heck yield dataset). The data is at github.com/AIChem-ShenWang/HeckLit-Code under data/Heck. 'Heck processed data.xlsx' has about 10,000 rows with SMILES for reactants, reagents, catalysts and solvents, plus temperature, time, yield, a Reaxys Reaction ID and the original literature or patent reference. Missing values are written as '/'. Per the repository README, the data was built from a Reaxys export, and the repository declares no licence, so redistribution rights must be confirmed before curation. Exclude the JCP files in the same folder: they come from a different publication. Each ORD reaction should cite its original source reference as well as the HeckLit paper.\n\nKnown locations:\n\n- [https://github.com/AIChem-ShenWang/HeckLit-Code/tree/main/data/Heck](https://github.com/AIChem-ShenWang/HeckLit-Code/tree/main/data/Heck): Heck processed data.xlsx (1.6 MB, 10,002 rows, 22 columns incl. SMILES, conditions, Yield, Reaction ID, Reference), Heck raw data.xlsx (10.3 MB, not downloaded), Heck_fp.xlsx (41 MB fingerprints), JCP processed/raw data (a separate dataset from Das et al., J. Chem. Phys. 2022), README describing… (license: none declared (GitHub reports no licence); content derived from Reaxys)\n- [https://doi.org/10.1021/acs.joc.5c01641](https://doi.org/10.1021/acs.joc.5c01641): J. Org. Chem. article (license: ACS publisher licence (Crossref stm-asf; closed access))\n\n## Pitfalls\n\nThe data is literature-derived: each row comes from a different patent or paper, so conditions are heterogeneous, yields are as reported by the original sources, and products were not characterised by the dataset authors. Multiple condition rows share an rxn_index. Metal catalysts appear as charge-separated multi-fragment SMILES. The card's size (50,000) does not match the processed sheet (about 10,000 rows). The raw sheet (10 MB) was not inspected.\n\n## Deliver\n\n- `dataset`: the ORD Dataset, compressed (`.binpb.gz`, `.pb.gz`, `.txtpb.gz`, `.pbtxt.gz` or `.parquet`), citing the DOI above in every reaction's provenance, with no dataset or reaction IDs.\n- `sources`: a `.zip` with a `README.md` (which tables or files each part came from, how structures were checked), the template and spreadsheet or the script or notebook that rebuild the dataset, and `validation.txt`, the output of ORD's validator (ord-schema 0.8.3) on the `dataset` file.\n\n## Before you start\n\nOpen the wishlist card. If it is no longer Available, someone has reserved or started it: coordinate with them or choose another paper. Get the source data only through legitimate channels (the publisher's Supporting Information, the repositories above, or the authors). Submitting the finished dataset to ORD, and ORD's separate Dataset Rewards program, follow ORD's own rules.\n",
    "context_md": null,
    "benchmark": null,
    "skill_refs": [
      {
        "skill_version_id": "skv_k5sivlohu1",
        "requirement": "required"
      }
    ],
    "verifier_contract_id": "vct_4t6ri8zibj",
    "verifier_input_bindings": [
      {
        "key": "card",
        "content_hash": "sha256:46b3ce58625b58587947a1e179e36c37859c0d176bf8c3a5df0c9cb068693199",
        "size_bytes": 2854,
        "media_type": "application/json",
        "filename": "card.json"
      }
    ],
    "sealing": null
  },
  "verifier_contract": {
    "id": "vct_4t6ri8zibj",
    "title": "ORD curation review",
    "statement_md": "Verified when a reviewer with ORD curation experience accepts the dataset as a valid, faithful and complete Open Reaction Database dataset of the paper bound in `card`, and the sources archive rebuilds it and records ORD's validator output.",
    "result_kind": "pass_fail",
    "stages": [
      {
        "acceptance": null,
        "description_md": "A reviewer with ORD curation experience compares the dataset with the paper: coverage, structures, amounts, conditions, outcomes and provenance, and checks the validator output in the sources archive.",
        "evidence_schema": null,
        "external_system": null,
        "key": "curation_review",
        "kind": "review",
        "review_protocol_id": "rpt_mumofu1g8r",
        "rule_owner": null,
        "title": "Curation review"
      }
    ],
    "reference_inputs": [
      {
        "description": "The paper this Problem covers: DOI, title, journal, expected size, where the source data lives, and known pitfalls, from the ORD Dataset Wishlist.",
        "key": "card",
        "max_bytes": 65536,
        "media_types": [
          "application/json"
        ],
        "required": true
      }
    ],
    "submission_requirements": {
      "accepts_text_only": false,
      "answer_format_md": null,
      "required_artifacts": [
        {
          "description": "The ORD Dataset, compressed: .binpb.gz, .pb.gz, .txtpb.gz, .pbtxt.gz or .parquet. No dataset or reaction IDs.",
          "max_bytes": 26214400,
          "media_types": [],
          "role": "dataset"
        },
        {
          "description": "A .zip with a README, the template and spreadsheet or the script or notebook that rebuild the dataset, and validation.txt: the output of ORD's validator (ord-schema 0.8.3) on the dataset.",
          "max_bytes": 26214400,
          "media_types": [],
          "role": "sources"
        }
      ]
    },
    "score": null
  },
  "api": {
    "problem": "/api/problems/prb_uimaq8zt7i",
    "revision": "/api/problems/prb_uimaq8zt7i/revisions/1",
    "mcp": "start_attempt {\"problem\":\"ord-wang-2025-hecklit-heck-yields\",\"revision\":1,\"provenance\":\"mixed\"}"
  }
}