{"id":"W4416414271","doi":"10.1038/s41597-025-06068-4","title":"The ICGC ARGO data dictionary for standardizing global cancer clinical data","year":2025,"lang":"en","type":"article","venue":"Scientific Data","topic":"Cancer Genomics and Diagnostics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Indoc Research; University of Toronto; Ontario Institute for Cancer Research","funders":"Government of Ontario","keywords":"Interoperability; Cancer; Clinical Oncology; Argo; Precision oncology; Germline; Genome; Genomics; Metadata","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["open_science"],"consensus_categories":["open_science"],"category_scores_codex":[0.002630925,0.0001103773,0.0001225045,0.00001501331,0.0006577027,0.0004328175,0.005604944,0.00008167856,0.000008817845],"category_scores_gemma":[0.001759383,0.00008791392,0.0000340817,0.0001821114,0.0004026971,0.00002351012,0.00962432,0.00007582407,0.000005672939],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000286944,"about_ca_system_score_gemma":0.001260458,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009102376,"about_ca_topic_score_gemma":0.002992562,"domain_scores_codex":[0.997802,0.00004253349,0.0003354289,0.001359043,0.0001780262,0.0002829276],"domain_scores_gemma":[0.9910634,0.0001434384,0.00009243779,0.008497741,0.0001306677,0.00007236462],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00008068014,0.00003578074,0.002062244,0.00001074542,0.00009292036,5.213252e-7,0.000001253274,0.00001221836,0.0004703285,0.0003863214,0.937451,0.05939596],"study_design_scores_gemma":[0.000411618,0.00002023199,0.001106842,0.00001921854,0.00007755686,0.000001101921,0.0000399333,0.004892235,0.0001806276,0.0004913928,0.9926544,0.0001047928],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.009116296,0.03493708,0.06291809,0.006839074,0.02567459,0.001037599,0.8572422,0.00003564155,0.002199451],"genre_scores_gemma":[0.1370176,0.02248385,0.01570608,0.003637727,0.006263277,0.0001250218,0.7999911,0.00007487826,0.0147004],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.1279013,"threshold_uncertainty_score":0.9997752,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1217371070011208,"score_gpt":0.456465764786295,"score_spread":0.3347286577851742,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}