{"id":"W1964226146","doi":"10.1109/bibm.2014.6999305","title":"Spaced seed data structures","year":2014,"lang":"en","type":"article","venue":"","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canada's Michael Smith Genome Sciences Centre","funders":"National Institutes of Health; Genome British Columbia; Genome Canada","keywords":"Sequence assembly; Hybrid genome assembly; k-mer; De Bruijn graph; Genome; Reference genome; Computer science; DNA sequencing; Bloom filter; Data structure; Genomics; Computational biology; Biology; Theoretical computer science; Algorithm; Graph; Genetics; DNA; Gene","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":[],"consensus_categories":[],"category_scores_codex":[0.00236331,0.001085551,0.001324589,0.002287165,0.001856428,0.003618546,0.004007101,0.001887889,0.01909436],"category_scores_gemma":[0.01039389,0.0009720553,0.00122852,0.003957901,0.001421852,0.007265109,0.003968623,0.002339712,0.008954629],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002080344,"about_ca_system_score_gemma":0.003004056,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002824973,"about_ca_topic_score_gemma":0.004385169,"domain_scores_codex":[0.9975705,0.0004753721,0.0005414246,0.0004744803,0.000704355,0.0002339017],"domain_scores_gemma":[0.9925929,0.002091806,0.0005281825,0.003003851,0.00148836,0.0002949387],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002105232,0.000360907,0.00410647,0.001369477,0.0001526677,0.0005541937,0.0007991928,0.02210815,0.01678495,0.3622589,0.1051627,0.4842372],"study_design_scores_gemma":[0.0004756745,0.0007884887,0.000783844,0.0004465317,0.00009702559,0.0009538222,0.0003417706,0.1400359,0.05352759,0.3729761,0.4293316,0.0002417162],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01763653,0.00154975,0.9267703,0.001407873,0.0006679334,0.0008593235,0.01256674,0.02265146,0.01589011],"genre_scores_gemma":[0.1388126,0.0009476499,0.8050425,0.001495408,0.0002570305,0.001787216,0.02757003,0.002546585,0.02154104],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01909436,"threshold_uncertainty_score":0.06387699,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01789176406345324,"score_gpt":0.2501607111455112,"score_spread":0.2322689470820579,"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."}}