{"id":"W4407114045","doi":"10.1186/s12859-025-06050-2","title":"BioLake: an RNA expression analysis framework for prostate cancer biomarker powered by data lakehouse","year":2025,"lang":"en","type":"article","venue":"BMC Bioinformatics","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Prostate Cancer Canada; University of Calgary","funders":"","keywords":"Upload; Computer science; Raw data; Scalability; Flexibility (engineering); Data science; Data pre-processing; Pipeline (software); Data mining; Workflow; Data warehouse; Database; World Wide Web","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":[0.004284732,0.002028242,0.001500122,0.002982506,0.001086354,0.003736529,0.004919149,0.001220463,0.01731568],"category_scores_gemma":[0.008798117,0.00167794,0.003222904,0.002541099,0.001155573,0.005517698,0.007396559,0.002800629,0.009032269],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001389906,"about_ca_system_score_gemma":0.002780239,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005959373,"about_ca_topic_score_gemma":0.01004063,"domain_scores_codex":[0.9973308,0.0003322803,0.0002884369,0.0007017556,0.001143952,0.0002027819],"domain_scores_gemma":[0.9969695,0.001072778,0.0002813405,0.0008718951,0.0005055767,0.0002989951],"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":[0.004318623,0.0004878395,0.01519864,0.00400788,0.001179337,0.002465914,0.002729213,0.03279492,0.0725819,0.08605979,0.453798,0.324378],"study_design_scores_gemma":[0.0008181265,0.0003543067,0.009185676,0.000509868,0.0003006234,0.001191009,0.0004790343,0.2453508,0.063432,0.1164494,0.5610933,0.0008359399],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.003875286,0.0006575483,0.5863896,0.0008191019,0.0002257076,0.0005780288,0.03127985,0.3718549,0.004320045],"genre_scores_gemma":[0.08307444,0.001224445,0.7035417,0.002366618,0.0002064509,0.003332761,0.1388619,0.05620665,0.01118496],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01731568,"threshold_uncertainty_score":0.05792671,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04424208771644974,"score_gpt":0.352946851264794,"score_spread":0.3087047635483443,"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."}}