AI and the World's Hard Problems · Medicine & drug discovery · Nature (computational biology) · 2026-10-23
Identification of complement system-related biomarkers in immunoglobulin A nephropathy through machine learning and single-cell data analysis
Complement dysregulation is linked to IgA nephropathy, but its molecular markers are not fully defined. The researchers combined bulk transcriptomic and single-cell RNA-seq datasets and used machine-learning feature selection, highlighting A2M, C3AR1 and SERPINA1 as candidate biomarkers. A three-gene neural network classifier reached an AUC of 0.959 on external validation, though overfitting remains possible and larger, clinically annotated cohorts are needed to confirm these markers.