{
  "abstract": "The Buchwald-Hartwig cross-coupling is a cornerstone of modern pharmaceutical synthesis, yet predictive modeling of its outcomes remains constrained by data quality and chemical space coverage. Electronic laboratory notebooks contain heterogeneous, noisy records, while open-source high-throughput experimentation (HTE) datasets are fragmented and narrow in scope, leading to poor model performance on unseen substrates and conditions. Here we introduce a framework that systematically standardizes and integrates multiple reaction datasets into a high-quality, unique-structure-per-entity dataset, coupled with active learning to strategically expand chemical space. By merging published Buchwald-Hartwig HTE data with new experimental results, we achieve a model with predictive power across novel substrates and conditions, delivering improved out-of-distribution predictions compared with previous approaches. Crucially, model-guided reagent recommendations were validated experimentally, confirming the framework's utility to uncover unexplored reactivity. This work establishes a blueprint for robust machine learning in synthetic chemistry and enables preemptive in silico reagent screening to accelerate pharmaceutical discovery.",
  "abstract_shown": true,
  "affiliations": [
    "Drug Discovery Data Science, In Silico Discovery, Johnson & Johnson, Porto Salvo, Portugal",
    "Laboratory of Artificial Chemical Intelligence (LIAC), Institut des Sciences et Ingénierie Chimiques, Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland",
    "Johnson & Johnson (United States)",
    "École Polytechnique Fédérale de Lausanne",
    "Chemistry Capabilities, Analytical and Purification, Global Discovery Chemistry, Johnson & Johnson, Spring House, PA, USA",
    "Drug Discovery Data Science, In Silico Discovery, Johnson & Johnson, Spring House, PA, USA",
    "Drug Discovery Data Science, In Silico Discovery, Johnson & Johnson, Cambridge, MA, USA. jwegner@its.jnj.com",
    "Laboratory of Artificial Chemical Intelligence (LIAC), Institut des Sciences et Ingénierie Chimiques, Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland. philippe.schwaller@epfl.ch",
    "National Centre of Competence in Research (NCCR) Catalysis, Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland. philippe.schwaller@epfl.ch",
    "Chemistry Capabilities, Analytical and Purification, Global Discovery Chemistry, Johnson & Johnson, Spring House, PA, USA. istrambe@its.jnj.com"
  ],
  "article_type": null,
  "arxiv": null,
  "author_total": 7,
  "authors": [
    {
      "affs": [
        0,
        1,
        2,
        3
      ],
      "name": "Paulo Neves",
      "orcid": "0000-0001-9556-6418"
    },
    {
      "affs": [
        4,
        2
      ],
      "name": "Bo Hao"
    },
    {
      "affs": [
        5,
        2
      ],
      "name": "Santeri Aikonen",
      "orcid": "0000-0003-2675-7290"
    },
    {
      "affs": [
        4,
        2
      ],
      "name": "Justin B. Diccianni",
      "orcid": "0009-0005-2922-6868"
    },
    {
      "affs": [
        6,
        2
      ],
      "name": "Jörg K. Wegner",
      "orcid": "0000-0002-1852-9434"
    },
    {
      "affs": [
        7,
        8,
        3
      ],
      "name": "Philippe Schwaller",
      "orcid": "0000-0003-3046-6576"
    },
    {
      "affs": [
        9,
        2
      ],
      "name": "Iulia I. Strambeanu",
      "orcid": "0000-0002-1502-5484"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/f0925q03d8625.bib",
  "cited_by": 5,
  "doi": "10.1038/s43588-026-01017-6",
  "fulltext_url": null,
  "id": "f0925q03d8625",
  "issns": [
    "2662-8457"
  ],
  "landing_url": "https://doi.org/10.1038/s43588-026-01017-6",
  "language": "en",
  "licence": "CC-BY",
  "licence_label": "CC BY",
  "licence_source": "journal",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [
    [
      "openalex",
      "3002|Computational Drug Discovery Methods"
    ]
  ],
  "openalex": "W4415161350",
  "pdf_url": "https://www.nature.com/articles/s43588-026-01017-6.pdf",
  "pmcid": "PMC13593570",
  "pmid": "42575946",
  "published": "2026-08-10",
  "publisher": "Nature Portfolio",
  "ris": "https://research.draconitic.com/paper/f0925q03d8625.ris",
  "sources": [
    {
      "first_seen": "2026-11-07T13:21:00Z",
      "native_id": "W4415161350",
      "source": "openalex",
      "url": "https://doi.org/10.1038/s43588-026-01017-6"
    }
  ],
  "subjects": [
    {
      "field": "chemistry",
      "origin": "openalex:3002|Computational Drug Discovery Methods",
      "subfield": "medicinal"
    }
  ],
  "title": "Robust out-of-distribution prediction of Buchwald–Hartwig reactions",
  "title_text": "Robust out-of-distribution prediction of Buchwald–Hartwig reactions",
  "type": "peer-reviewed",
  "updated_at": "2026-11-09T13:21:17Z",
  "url": "https://research.draconitic.com/paper/f0925q03d8625",
  "venue": "Nature Computational Science",
  "versions": []
}