Population-scale genomics reveals divergent pathogenicity of variant classes across paralogous collagen IV genes

Monoallelic pathogenic or likely pathogenic variants in COL4A3 and COL4A4 occur in approximately 1 in 106 individuals, yet whether these paralogous genes confer equivalent pathogenicity for the same variant classes has not been tested at population scale. Using whole-genome sequencing data from the UK Biobank (UKB; n ≈ 500,000), with replication in the All of Us Research Program (n ≈ 414,000), we performed per-variant association testing, gene-based collapsing analyses, and phenome-wide association studies (PheWASs) across hematuria, proteinuria, and chronic kidney disease. We identified 64 COL4A3 and 92 COL4A4 rare variants significantly associated with hematuria or proteinuria, providing population-level quantitative evidence that may support variant interpretation. Glycine substitutions within collagenous domains conferred similar risks in both genes. In contrast, truncating and non-collagenous domain (NC1) missense variants were strongly associated with hematuria and proteinuria in COL4A4 heterozygotes but showed substantially attenuated or absent associations in COL4A3 heterozygotes despite comparable allele frequencies and predicted pathogenicity scores. These findings were independently replicated in All of Us. Genome-wide association analysis identified the COL4A3 and COL4A4 locus as the dominant genetic determinant of hematuria, attributable to the aggregate effects of rare coding variants with no evidence of independent common variant or trans-acting modifier effects. These findings demonstrate marked gene-specific differences in association with hematuria and proteinuria for truncating and NC1 variants between COL4A3 and COL4A4, challenging assumptions of equivalent pathogenicity across paralogous collagen IV genes. Gene identity, not variant class alone, should inform risk stratification, variant interpretation, and genetic counseling in collagen IV heterozygotes.