{"id":"W4400014999","doi":"10.1101/2024.06.21.599698","title":"A roadmap for fair reuse of public microbiome data","year":2024,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Reuse; Microbiome; Data science; Business; World Wide Web; Computer science; Biology; Ecology; Bioinformatics","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":[0.5670183,0.002667441,0.003218997,0.01251738,0.009361664,0.03799465,0.01419121,0.02468662,0.02002222],"category_scores_gemma":[0.5940855,0.004717319,0.006640094,0.00955926,0.02430736,0.07094985,0.05374885,0.03887607,0.02064826],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0119167,"about_ca_system_score_gemma":0.1133868,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009299608,"about_ca_topic_score_gemma":0.006497583,"domain_scores_codex":[0.4545327,0.3385529,0.05824513,0.02865447,0.10318,0.01683489],"domain_scores_gemma":[0.2179184,0.2877861,0.02041159,0.2524714,0.1778369,0.04357573],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0004631967,0.001160345,0.004631183,0.002701455,0.0002384959,0.0007902087,0.008219612,0.002645801,0.003780687,0.3403066,0.2802962,0.3547663],"study_design_scores_gemma":[0.0003230096,0.0004278133,0.00239046,0.004856874,0.0001019929,0.0006284809,0.00409485,0.003652282,0.003759463,0.2479626,0.731416,0.0003859987],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.004691656,0.004214386,0.5475647,0.386223,0.008495484,0.006025619,0.001612665,0.008270787,0.03290161],"genre_scores_gemma":[0.02481215,0.002978261,0.8967345,0.04756632,0.002721629,0.007590017,0.004342151,0.003797831,0.009457187],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9858088,"threshold_uncertainty_score":0.5339436,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.3519546371755455,"score_gpt":0.4571217903180451,"score_spread":0.1051671531424996,"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."}}