Canonical correlation analysis identifies single nucleotide polymorphism combinations and gene expression associated with membranous nephropathy

Abstract In membranous nephropathy (MN), an autoimmune kidney condition, univariate genome-wide association studies (GWAS) have identified Human Leucocyte Antigen ( HLA ) and M-type phospholipase A2 receptor ( PLA2R1 ) genetic risk variants. Multivariate-single nucleotide polymorphism (SNP) canonical correlation analysis (CCA) with transcriptomics analysis, can uncover additional insights within this context. We applied CCA to our National Unified Renal Translational Research Enterprise (NURTuRE)- chronic kidney disease (CKD) and Salford Kidney Study (SKS) European datasets, identifying 1095 HLA and 46 PLA2R1 SNPs showing significant MN correlation, and single-cell gene enrichment. Top SNP combinations showed greater significance in both MN by multivariate-SNP CCA, and MN and diabetic nephropathy (Type 1) by allele frequency analysis compared to healthy and non-MN groups. ATF6B , C2, CYP21A2 , and HLA-DRB1 colocalised with top HLA SNPs and showed significant differential expression in MN (published immune transcriptomics datasets). HLA-DQB1, HLA-DRB5 and RNF5 showed significant differential expression with lead SNP genotype status in MN and CKD. These previously unreported SNP and gene results are candidates for MN and potentially Type 1 diabetic nephropathy risk prediction.