{"schema_version":"1.0","hypothesis_id":39232,"triple":{"drug":"tocilizumab","drug_canonical":"tocilizumab","mediator":"C-reactive protein","mediator_canonical":"c-reactive protein","outcome":"type 2 diabetes","outcome_canonical":"type 2 diabetes"},"domain":{"domain":"cross_domain","source":"rheumatoid_arthritis","target":"type_2_diabetes"},"provenance_completeness":"partial_backfill","completeness_notes":"Backfilled retroactively from DB state on 2026-05-13. The pipeline_run this row references was created by the backfill script, not by the actual generation orchestrator. Source papers, claims, scoring, and embedding config are recovered; generation-time metadata (exact git commit at gen time, host, exact command) is not recoverable.","source_papers":[{"doi":"10.1016/j.esmoop.2024.104088","pmid":"39754984","year":2025,"roles":["drug_mediator"],"title":"Anti-IL-6R Ab tocilizumab to treat paraneoplastic inflammatory syndrome of solid cancers","journal":"ESMO Open","work_id":288301,"publication_date":"2025-01-01"},{"doi":"10.1038/s41598-019-40987-8","pmid":"30872696","year":2019,"roles":["mediator_outcome"],"title":"Association of C-Reactive Protein with Risk of Developing Type 2 Diabetes Mellitus, and Role of Obesity and Hypertension: A Large Population-Based Korean Cohort Study","journal":"Sci Rep","work_id":235329,"publication_date":"2019-03-14"},{"doi":"10.1186/s12933-016-0369-6","pmid":"27013319","year":2016,"roles":["mediator_outcome"],"title":"Oxidative stress, inflammation, endothelial dysfunction and incidence of type 2 diabetes","journal":"Cardiovasc Diabetol","work_id":232291,"publication_date":"2016-03-24"}],"source_paper_count":3,"extracted_claims":[{"side":"drug_mediator","object":"C-reactive protein","context":"in patients with solid cancers and severe PIS","subject":"tocilizumab","work_id":288301,"claim_id":212704,"polarity":"positive","predicate":"downregulates","confidence":0.8,"extraction_run_id":29},{"side":"mediator_outcome","object":"type 2 diabetes","context":"in the CARDIA study, after full adjustment","subject":"C-reactive protein","work_id":232291,"claim_id":164888,"polarity":"positive","predicate":"correlates_with","confidence":0.95,"extraction_run_id":27},{"side":"mediator_outcome","object":"type 2 diabetes","context":"in a large population-based Korean cohort","subject":"C-reactive protein","work_id":235329,"claim_id":172812,"polarity":"positive","predicate":"associated_with","confidence":0.9,"extraction_run_id":27},{"side":"mediator_outcome","object":"type 2 diabetes","context":"in women (highest CRP tertile versus lowest, HR 2.80)","subject":"C-reactive protein","work_id":235329,"claim_id":172813,"polarity":"positive","predicate":"associated_with","confidence":0.85,"extraction_run_id":27},{"side":"mediator_outcome","object":"type 2 diabetes","context":"in men (highest CRP tertile versus lowest, HR 1.67)","subject":"C-reactive protein","work_id":235329,"claim_id":172814,"polarity":"positive","predicate":"associated_with","confidence":0.85,"extraction_run_id":27}],"extracted_claim_count":5,"extraction":{"model":"deepseek-v4-flash","model_source":"configured","prompt_version":"v1_2026_05_07","run_ids":[27,29]},"bridge":{"a_to_b_predicate":"downregulates","b_to_c_predicate":"correlates_with","a_to_b_evidence_count":1,"b_to_c_evidence_count":4,"edge_support_log_ab":0.6931471805599453,"edge_support_log_bc":1.6094379124341003,"novelty_factor":1.0,"direct_a_to_c_count":0},"embedding":{"model":"bge-m3","dim":1024,"max_length":8192,"similarity_scores":{}},"scoring":{"method":"swanson_abc_log_evidence_with_novelty_penalty","formula":"score = min((log(ab+1) + log(bc+1)) * novelty / 10, 1.0); novelty = 1.0 if direct==0 else 0.7 / (direct+1)","score_raw":0.23025850929940456,"score":0.23025850929940456,"quality_tier":"mid","quality_reason":null},"pipeline_run":{"id":4,"run_uuid":"eb8b7730-d92e-4cd5-95dd-95f6c5d62cc2","command":"scripts/backfill_provenance_v2.py:v2_2026_05_12","status":"completed","git_commit":"9996ffaa56e9b7cd5f677bd8203624c2c18aa000","host":"MacBook-Pro-Daniel96.local","started_at":"2026-05-12T21:35:53.363326+00:00","ended_at":"2026-05-12T21:43:41.903387+00:00","notes":"Retroactive provenance backfill on 2026-05-13 for snapshot v2_2026_05_12. Synthetic pipeline_runs row — original generation run records are lost (pre-date provenance schema). All recoverable fields populated from current DB state (claims, works, repurposing_hypotheses); pipeline_run record is synthetic, hence completeness='partial_backfill'.\nBackfilled 1,618 provenance rows. Source papers: avg=3.73, zero=0. Claims matched: avg=4.09, zero=0.\nRefreshed 1,618 provenance rows. Source papers: avg=3.73, zero=0. Claims matched: avg=4.09, zero=0.\nRefreshed 1,618 provenance rows. Source papers: avg=3.73, zero=0. Claims matched: avg=4.09, zero=0.","domains":["als","alzheimer","depression","epilepsy","glioblastoma","inflammatory_bowel_disease","multiple_sclerosis","pancreatic_cancer","rheumatoid_arthritis","type_2_diabetes"]},"generated_at":"2026-05-12T21:43:40.945575+00:00"}