Identifying functional subtypes of IgA nephropathy based on three machine learning algorithms and WGCNA

Area of Science:

• Nephrology
• Immunology
• Genomics

Background:

• IgA nephropathy (IgAN) is a leading cause of renal failure, but its molecular mechanisms and classification remain limited.
• Current IgAN classification relies heavily on pathology, lacking molecular-level insights into disease heterogeneity.

Purpose of the Study:

• To identify molecular subtypes of IgA nephropathy (IgAN).
• To explore the functional and immune cell infiltration heterogeneity across identified IgAN subtypes.
• To establish a classifier for IgAN subtypes to aid in future precise treatments.

Main Methods:

• Utilized microarray datasets (GSE116626, GSE115857) to identify differentially expressed genes (DEGs) in IgAN.
• Employed unsupervised clustering algorithms (hclust, PAM, ConsensusClusterPlus) and random forest to classify IgAN subtypes.
• Applied Weighted Gene Co-expression Network Analysis (WGCNA) and CIBERSORT for immune cell infiltration analysis to verify subtype heterogeneity.

Main Results:

• Identified 102 DEGs and three distinct functional IgAN subtypes: viral-hormonal, bacterial-immune, and mixed.
• Discovered subtype-specific marker genes (e.g., ATF3, JUN, FOS for viral-hormonal; LIF, C19orf51, SLPI for bacterial-immune) correlated with proteinuria and eGFR.
• Confirmed subtype heterogeneity through WGCNA modules and distinct immune cell infiltration patterns (e.g., T follicular helper cells, macrophages, eosinophils).

Conclusions:

• Successfully identified three novel functional subtypes of IgA nephropathy (IgAN) at the molecular level.
• Developed a subtype classifier with potential for precise treatment strategies in IgAN patients.
• Demonstrated significant molecular and immune heterogeneity among IgAN subtypes.