{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-08-07-026",
    "source_story_id": "tmp-story-metabollm-metabolite-graphs",
    "edition_id": "mp-2026-08-07-morning-0029",
    "edition_url": "https://themachinepress.com/edition/2026-08-07",
    "position": 15,
    "story_type": "dispatch",
    "section": "research",
    "editorial_classification": "editorial",
    "headline": "Biochemical Prose Became a Patient-Level Graph",
    "slug": "biochemical-prose-became-a-patient-level-graph",
    "dek": "MetaboLLM turns retrieved metabolomics descriptions into graph structures used for two downstream prediction tasks.",
    "summary": "MetaboLLM turns retrieved metabolomics descriptions into graph structures used for two downstream prediction tasks.",
    "body_text": "MetaboLLM combines continual pretraining, supervised tuning and structured retrieval, then converts its biochemical descriptions into metabolite graphs for a graph neural network. The authors report AUCs of 0.8616 for stress hyperglycemia after coronary bypass and 0.8123 for postmenopausal hormone-regimen classification, ahead of their tested alternatives. These retrospective benchmark results do not establish prospective clinical utility or causal biochemical mechanisms.",
    "why_it_matters": "MetaboLLM turns retrieved metabolomics descriptions into graph structures used for two downstream prediction tasks.",
    "limitations": [
      "These retrospective benchmark results do not establish prospective clinical utility or causal biochemical mechanisms."
    ],
    "importance": 7,
    "canonical_url": "https://themachinepress.com/story/mp-2026-08-07-026/biochemical-prose-became-a-patient-level-graph",
    "json_url": "https://themachinepress.com/story/mp-2026-08-07-026.json",
    "first_published_at": "2026-08-07T09:00:00.000-04:00",
    "modified_at": "2026-08-07T09:00:00.000-04:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-08-07-026-001",
        "text": "MetaboLLM turns retrieved metabolomics descriptions into graph structures used for two downstream prediction tasks.",
        "source_ids": [
          "source-2026-08-07-015"
        ],
        "qualification": "These retrospective benchmark results do not establish prospective clinical utility or causal biochemical mechanisms."
      }
    ],
    "source_ids": [
      "source-2026-08-07-015"
    ],
    "tags": [
      "metabolomics",
      "language models",
      "graph learning"
    ],
    "image_url": null,
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-08-07-015",
      "title": "arXiv preprint 2608.06253",
      "publisher": "arXiv",
      "url": "https://arxiv.org/abs/2608.06253",
      "canonical_url": "https://arxiv.org/abs/2608.06253",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": null,
      "accessed_at": "2026-08-07T08:22:48.000-04:00",
      "supports_claim_ids": [
        "claim-mp-2026-08-07-026-001"
      ]
    }
  ],
  "corrections": [],
  "publisher": {
    "name": "The Machine Press",
    "url": "https://themachinepress.com",
    "description": "A daily newspaper for the age of artificial intelligence."
  },
  "cite_this_report": {
    "title": "Biochemical Prose Became a Patient-Level Graph",
    "publisher": "The Machine Press",
    "published_at": "2026-08-07T09:00:00.000-04:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-08-07-026/biochemical-prose-became-a-patient-level-graph"
  }
}