{"id":"W4280519287","doi":"10.12688/f1000research.109080.2","title":"Recommendations for the formatting of Variant Call Format (VCF) files to make plant genotyping data FAIR","year":2022,"lang":"en","type":"preprint","venue":"F1000Research","topic":"Research Data Management Practices","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"ca_institutions":"Ontario Institute for Cancer Research","funders":"Biotechnology and Biological Sciences Research Council; Horizon 2020 Framework Programme; German Network for Bioinformatics Infrastructure; Bundesministerium für Bildung und Forschung; European Commission","keywords":"Disk formatting; Metadata; Computer science; Information retrieval; Identifier; World Wide Web; Programming language","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":[0.1156209,0.001552465,0.001413872,0.007683923,0.003114025,0.01101377,0.007959653,0.01202422,0.06379459],"category_scores_gemma":[0.3252777,0.002406436,0.003023155,0.007559527,0.005030325,0.01694901,0.006880476,0.01232317,0.06312889],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.006044528,"about_ca_system_score_gemma":0.01960644,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01820415,"about_ca_topic_score_gemma":0.01250271,"domain_scores_codex":[0.9193949,0.03804471,0.01664544,0.003213192,0.01998881,0.002712899],"domain_scores_gemma":[0.6098298,0.1514472,0.02022919,0.07010361,0.1383874,0.01000275],"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":[0.0002705881,0.0002456038,0.001795676,0.001245948,0.00004140812,0.0005304182,0.002518971,0.002083395,0.004613827,0.03365608,0.799081,0.1539171],"study_design_scores_gemma":[0.0001085797,0.0000764429,0.001187069,0.002126287,0.00002938361,0.0002971202,0.0006894265,0.001111326,0.003264847,0.01456923,0.9763659,0.0001744024],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.004965193,0.003342097,0.6126712,0.1426275,0.01880666,0.01221495,0.03153472,0.04588748,0.1279502],"genre_scores_gemma":[0.01213853,0.003413033,0.7893967,0.05670411,0.002799331,0.01177082,0.03336461,0.01940689,0.07100597],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9920403,"threshold_uncertainty_score":0.6114693,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4089725583043259,"score_gpt":0.4605710201384414,"score_spread":0.05159846183411543,"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."}}