{"id":"W2335458838","doi":"10.1038/sdata.2014.12","title":"Direct infusion mass spectrometry metabolomics dataset: a benchmark for data processing and quality control","year":2014,"lang":"en","type":"article","venue":"Scientific Data","topic":"Metabolomics and Mass Spectrometry Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":139,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"Pfizer Canada; University of Birmingham; British Heart Foundation; Natural Environment Research Council; Advantage West Midlands; Pfizer","keywords":"Workflow; Benchmark (surveying); Computer science; Metabolomics; Data mining; Quality (philosophy); Set (abstract data type); Data quality; Quality assurance; Bioinformatics; Biology; Database; Medicine; Engineering","routes":{"ca_aff":true,"ca_fund":true,"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"],"consensus_categories":[],"category_scores_codex":[0.007343716,0.001738977,0.001361757,0.002837037,0.001269182,0.002420684,0.003499327,0.002392884,0.002201657],"category_scores_gemma":[0.01443356,0.0004473137,0.001451814,0.004072123,0.0008612641,0.001195008,0.002717301,0.001750336,0.003201236],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001550394,"about_ca_system_score_gemma":0.002940295,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005543026,"about_ca_topic_score_gemma":0.006738026,"domain_scores_codex":[0.9944722,0.001100308,0.0007226783,0.001399696,0.002021573,0.0002835328],"domain_scores_gemma":[0.9892395,0.002396666,0.001003084,0.003027365,0.003785752,0.000547686],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.006588144,0.002284706,0.06561458,0.00631315,0.002089899,0.001394029,0.0004873082,0.05296252,0.06338112,0.006054839,0.5819332,0.2108965],"study_design_scores_gemma":[0.001783772,0.00177169,0.1286825,0.0007086119,0.0006286683,0.002827784,0.0005034279,0.1147002,0.07089192,0.01598367,0.6609921,0.0005256102],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.1209397,0.00267349,0.05191797,0.002017451,0.0004811721,0.001328448,0.8020661,0.01286498,0.00571064],"genre_scores_gemma":[0.04075873,0.0004315121,0.05643514,0.0003432924,0.00006974349,0.0009454638,0.8996536,0.0006704089,0.0006921199],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.9926563,"threshold_uncertainty_score":0.03883773,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04716827421050954,"score_gpt":0.3218757350716982,"score_spread":0.2747074608611887,"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."}}