{"id":"W6949619380","doi":"10.5281/zenodo.15241443","title":"Opening up translational data impact through the Data Citation Corpus","year":2025,"lang":"en","type":"article","venue":"Zenodo (CERN European Organization for Nuclear Research)","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"OpenAlex","funders":"National Center for Advancing Translational Sciences","keywords":"Metadata; Citation; Identifier; Identification (biology); Genomics; Translational research; Profiling (computer programming); Unique identifier; Data curation","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch","open_science"],"consensus_categories":[],"category_scores_codex":[0.02218647,0.00108415,0.001900608,0.1072141,0.002859328,0.01209139,0.002369797,0.001880642,0.02100861],"category_scores_gemma":[0.1886228,0.0008935108,0.001148683,0.1405372,0.002210974,0.007883946,0.01016458,0.002598206,0.007645454],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00461273,"about_ca_system_score_gemma":0.008012157,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01116096,"about_ca_topic_score_gemma":0.01474179,"domain_scores_codex":[0.968379,0.006687012,0.008588947,0.004594446,0.01051387,0.001236678],"domain_scores_gemma":[0.7641327,0.1649597,0.02292824,0.01980516,0.02549079,0.002683253],"domain_codex":null,"domain_gemma":"reproducibility","domain_candidate":"reproducibility","domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0004054736,0.0001507498,0.0423201,0.01255171,0.0004029356,0.000636978,0.007750018,0.003103931,0.00301466,0.08634988,0.6722465,0.1710671],"study_design_scores_gemma":[0.00009267996,0.00004333676,0.03701151,0.002930925,0.0001528247,0.0001975915,0.002685886,0.002222434,0.001756343,0.02430286,0.9284346,0.0001689812],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"methods","genre_scores_codex":[0.01955905,0.003323612,0.01081564,0.003299887,0.0007587041,0.0005940536,0.9397188,0.002670575,0.01925957],"genre_scores_gemma":[0.06315193,0.003052796,0.04810686,0.0007997937,0.0006903686,0.004698036,0.8731878,0.002248826,0.004063567],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9976302,"threshold_uncertainty_score":0.1173347,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1238279532537489,"score_gpt":0.3570606096248361,"score_spread":0.2332326563710872,"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."}}