{"id":"W2898449026","doi":"10.1093/bioinformatics/bty888","title":"Alignment-free clustering of UMI tagged DNA molecules","year":2018,"lang":"en","type":"article","venue":"Bioinformatics","topic":"Cancer Genomics and Diagnostics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":44,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia; Simon Fraser University","funders":"","keywords":"Barcode; Cluster analysis; Computer science; DNA sequencing; Computational biology; Genomics; Data mining; Biology; DNA; Genetics; Genome; Gene; Artificial intelligence","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":[],"consensus_categories":[],"category_scores_codex":[0.001063631,0.001384048,0.001310248,0.002249307,0.001468027,0.001597849,0.003877799,0.001417844,0.004059702],"category_scores_gemma":[0.004792375,0.0006795187,0.00134223,0.003035612,0.00093043,0.001203833,0.001444735,0.001244261,0.003095842],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001853607,"about_ca_system_score_gemma":0.001995431,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01151009,"about_ca_topic_score_gemma":0.01359655,"domain_scores_codex":[0.9987215,0.0002068213,0.00006294881,0.0004860449,0.0004198805,0.0001027063],"domain_scores_gemma":[0.9980848,0.0005556251,0.0002095057,0.0003492073,0.0006772246,0.0001236166],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0006706452,0.0001675806,0.006443489,0.0008936186,0.0002534208,0.0004095414,0.0004310145,0.7523385,0.04344477,0.01157111,0.01443783,0.1689385],"study_design_scores_gemma":[0.00002745724,0.00004090067,0.0008223235,0.00002040608,0.00001944217,0.00006793826,0.00005210015,0.9687345,0.01826582,0.008170489,0.003753654,0.00002501371],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.04201275,0.0004067015,0.9412718,0.0001895558,0.00009699494,0.0002437953,0.001882527,0.01111202,0.002783765],"genre_scores_gemma":[0.1695755,0.0003162551,0.8118967,0.0002217136,0.00004553792,0.0004288077,0.01235898,0.001716353,0.003440209],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01151009,"threshold_uncertainty_score":0.02288622,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.00933349121835504,"score_gpt":0.2314702039192874,"score_spread":0.2221367127009323,"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."}}