Plasma proteome profiling identifies novel biomarkers and predictors for schizophrenia - PubMed
4 hours ago
- #Plasma Proteomics
- #Schizophrenia Biomarkers
- #Predictive Modeling
- Study utilized proteomic and longitudinal data from UK Biobank (70 schizophrenia cases, 36,135 controls) to identify protein biomarkers linked to schizophrenia onset.
- Identified 31 significant schizophrenia-related proteins via Cox regressions, with five showing non-linear associations; Lasso-Cox regression highlighted 17 proteins with non-zero coefficients.
- Mendelian randomization indicated JAM3 has a positive causal relationship with schizophrenia (β = 0.1459, P = 0.0043).
- LightGBM model combining protein and population features demonstrated strong predictive ability (AUC = 0.7930, G-mean = 0.6683).
- Enrichment analyses revealed 31 proteins are involved in biological functions like cell adhesion, integrin binding, chemotaxis, and extracellular matrix organization.