{"id":"W3127575774","doi":"10.1101/2021.01.31.429071","title":"A topological characterization of DNA sequences based on chaos geometry and persistent homology","year":2021,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"Fractal and DNA sequence analysis","field":"Biochemistry, Genetics and Molecular Biology","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Topological data analysis; Persistent homology; Topology (electrical circuits); Homology (biology); DNA sequencing; Genome; Computational biology; DNA; Genetics; Biology; Mathematics; Algorithm; Gene; Combinatorics","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":[0.0004350786,0.0003654494,0.0005229989,0.007925211,0.0004657284,0.001439668,0.0005980825,0.0005371937,0.00103767],"category_scores_gemma":[0.002897897,0.0002138676,0.0005378521,0.002988619,0.001306953,0.001923705,0.001226339,0.0005267652,0.000280035],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006172915,"about_ca_system_score_gemma":0.0005256505,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00107815,"about_ca_topic_score_gemma":0.0008536995,"domain_scores_codex":[0.9994113,0.0001109294,0.00004744381,0.0001489644,0.0002100304,0.00007136934],"domain_scores_gemma":[0.9976552,0.0006483175,0.0006534466,0.0003191808,0.0004796861,0.0002442853],"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":[0.0004898183,0.000187244,0.04146934,0.0007158366,0.000185167,0.0009572789,0.001376048,0.1836891,0.1899024,0.2587205,0.00476029,0.3175469],"study_design_scores_gemma":[0.00003720459,0.0002725886,0.01879662,0.00005958583,0.00005109732,0.001249915,0.0005515988,0.8566436,0.02757843,0.08656456,0.008098002,0.00009677011],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.289147,0.0004422728,0.7060626,0.0002798408,0.00004310947,0.00009062287,0.0009705388,0.000654531,0.002309507],"genre_scores_gemma":[0.8522004,0.0002622444,0.1452583,0.00004325748,0.00007293914,0.00008168976,0.001133105,0.00008883802,0.0008591902],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007925211,"threshold_uncertainty_score":0.004478812,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.011044737686871,"score_gpt":0.2116497087783826,"score_spread":0.2006049710915116,"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."}}