{"id":"W4415526176","doi":"10.1101/2025.10.21.25338483","title":"Enriching for Answers in Rare Diseases","year":2025,"lang":"en","type":"preprint","venue":"medRxiv","topic":"Genomics and Rare Diseases","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Columbia College","funders":"National Institutes of Health","keywords":"Sequence (biology); Sampling (signal processing); Scalability; Sample (material); Phaser; Fraction (chemistry); DNA sequencing","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":[0.002063837,0.0006810669,0.0007209558,0.0008588523,0.0004218205,0.001489884,0.0007564797,0.0009962284,0.005449778],"category_scores_gemma":[0.005040167,0.0004795634,0.0005172692,0.0005964422,0.0005677882,0.0009491846,0.001935098,0.001647922,0.003365362],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003657436,"about_ca_system_score_gemma":0.0008688871,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005932223,"about_ca_topic_score_gemma":0.001339193,"domain_scores_codex":[0.998609,0.0003294993,0.00006372898,0.0004033413,0.0004736435,0.0001207623],"domain_scores_gemma":[0.9975648,0.001333349,0.0001804464,0.0004387723,0.0002969851,0.0001857877],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0006825493,0.00006665901,0.005413983,0.0008892473,0.0001973955,0.0003640023,0.0004625469,0.004262956,0.7149561,0.01631298,0.01225701,0.2441346],"study_design_scores_gemma":[0.0001441458,0.0003212719,0.005495455,0.000138093,0.0001724195,0.001075643,0.0001722703,0.04042389,0.7667418,0.03597476,0.1492335,0.0001068132],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1857102,0.009889008,0.7630984,0.005493996,0.001169343,0.0002686363,0.0109222,0.01397571,0.009472507],"genre_scores_gemma":[0.4186371,0.004434555,0.5431798,0.004227076,0.0006149092,0.0004401359,0.01601917,0.002171512,0.01027578],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005449778,"threshold_uncertainty_score":0.01823133,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01006658187027264,"score_gpt":0.2689616179235297,"score_spread":0.2588950360532571,"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."}}