{"id":"W4411852564","doi":"10.1101/2025.06.27.25330344","title":"Benchmarking the AI-based diagnostic potential of plasma proteomics for neurodegenerative disease in 17,187 people","year":2025,"lang":"en","type":"preprint","venue":"medRxiv","topic":"Alzheimer's disease research and treatments","field":"Medicine","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal; Institut Universitaire de Gériatrie de Montréal","funders":"Parkinsonfonden; Science for Life Laboratory; Skånes universitetssjukhus; Konung Gustaf V:s och Drottning Victorias Frimurarestiftelse; Vetenskapsrådet; Lunds Universitet; Knut och Alice Wallenbergs Stiftelse; Australian Government; GHR Foundation; Alzheimer's Association","keywords":"Proteomics; Biomarker; Disease; Medicine; Diagnostic accuracy; Dementia; Differential diagnosis; Bioinformatics; Pathology; Computational biology; Internal medicine; Biology","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":[0.001966881,0.001071478,0.0005184421,0.0005667677,0.0002065664,0.0008509735,0.0004835157,0.000796788,0.0009674632],"category_scores_gemma":[0.003524961,0.0002193946,0.0007175704,0.0003013521,0.0004272075,0.0004528413,0.0007553793,0.00070691,0.0006092919],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005532396,"about_ca_system_score_gemma":0.0005738608,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005394012,"about_ca_topic_score_gemma":0.004854719,"domain_scores_codex":[0.9995117,0.0001627441,0.000027235,0.0001938021,0.00005285563,0.00005155393],"domain_scores_gemma":[0.9993844,0.0002953054,0.00005650302,0.0001033994,0.00009613422,0.00006426348],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.004401599,0.0009591383,0.6100142,0.0004984557,0.002657186,0.001239161,0.0005533107,0.1436534,0.01468419,0.001407433,0.01934109,0.2005909],"study_design_scores_gemma":[0.0003356464,0.00154455,0.1782576,0.0001232766,0.000928122,0.001788669,0.0004079156,0.7819904,0.01913635,0.008109937,0.007260375,0.0001171618],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9761952,0.001339809,0.01476663,0.0006699798,0.0001129123,0.00006383644,0.004187498,0.001012119,0.001652158],"genre_scores_gemma":[0.9893591,0.0002379173,0.005122401,0.0001494587,0.00003240932,0.00002477418,0.004340395,0.00003340531,0.0007001032],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.005394012,"threshold_uncertainty_score":0.01072526,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02134648486946551,"score_gpt":0.3104948458782391,"score_spread":0.2891483610087736,"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."}}