{"id":"W2148425737","doi":"10.1093/bioinformatics/btu558","title":"BioBloom tools: fast, accurate and memory-efficient host species sequence screening using bloom filters","year":2014,"lang":"en","type":"article","venue":"Bioinformatics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":134,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canada's Michael Smith Genome Sciences Centre","funders":"Genome British Columbia; Genome Canada","keywords":"Bloom filter; False positive paradox; Computer science; Set (abstract data type); Sequence (biology); Filter (signal processing); Host (biology); Software; False positives and false negatives; Data set; Data structure; Data mining; Algorithm; Artificial intelligence; Programming language; Biology","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":[0.004967114,0.001787595,0.001670564,0.004585905,0.001613685,0.002734296,0.002977679,0.001490534,0.01999735],"category_scores_gemma":[0.01412492,0.001672868,0.00116091,0.002989664,0.0007670175,0.003995651,0.002906823,0.001698318,0.01774834],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001336472,"about_ca_system_score_gemma":0.002204797,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002277156,"about_ca_topic_score_gemma":0.003021455,"domain_scores_codex":[0.9979094,0.0004568511,0.000219338,0.0004800123,0.0007746591,0.0001596526],"domain_scores_gemma":[0.9960678,0.002282119,0.0004430407,0.0004538638,0.0004985417,0.0002545969],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002949911,0.00040296,0.01490312,0.003303602,0.000499701,0.0007779607,0.001198123,0.007655056,0.220705,0.01637815,0.3658251,0.3654013],"study_design_scores_gemma":[0.0008943445,0.0004354275,0.01216384,0.0007007049,0.0002256444,0.001581118,0.0004227062,0.2489151,0.3259519,0.04803422,0.3601913,0.0004836101],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.02390108,0.001228352,0.6466751,0.0006451826,0.0002506405,0.0005957832,0.03549718,0.2855703,0.005636415],"genre_scores_gemma":[0.06936495,0.0005946317,0.8580329,0.0005197127,0.00009345372,0.001506423,0.0473386,0.01676928,0.005780076],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.01999735,"threshold_uncertainty_score":0.06689781,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04576000212109506,"score_gpt":0.2551211840184869,"score_spread":0.2093611818973918,"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."}}