Evaluating Mendelian nephrotic syndrome genes for evidence of risk alleles or oligogenicity that explain heritability

Abstract

BackgroundMore than 30 genes can harbor rare exonic variants sufficient to cause nephrotic syndrome (NS), and the number of genes implicated in monogenic NS continues to grow. However, outside the first year of life, the majority of affected patients, particularly in ancestrally mixed populations, do not have a known monogenic form of NS. Even in those children classified with a monogenic form of NS, there is phenotypic heterogeneity. Thus, we have only discovered a fraction of the heritability of NS - the underlying genetic factors contributing to phenotypic variation. Part of the \"missing heritability\" for NS has been posited to be explained by patients harboring coding variants across one or more previously implicated NS genes, insufficient to cause NS in a classical Mendelian manner, but that nonetheless impact protein function enough to cause disease. However, systematic evaluation in patients with NS for rare or low-frequency risk alleles within single genes, or in combination across genes (\"oligogenicity\"), has not been reported.\n\nObjectiveTo determine whether, as compared to a reference population, patients with NS have either a significantly increased burden of protein-altering variants (\"risk-alleles\"), or unique combination of them (\"oligogenicity\"), in a set of 21 genes implicated in Mendelian forms of NS.\n\nMethodsIn 303 patients with NS enrolled in the Nephrotic Syndrome Study Network (NEPTUNE), we performed targeted amplification paired with next-generation sequencing of 21 genes implicated in monogenic NS. We created a high-quality variant call set and compared it to a variant call set of the same genes in a reference population composed of 2535 individuals from Phase 3 of 1000 Genomes Project. We created both a \"stringent\" and \"relaxed\" pathogenicity filtering pipeline, applied them to both cohorts, and computed the (1) burden of variants in the entire gene set per cohort, (2) burden of variants in the entire gene set per individual, (3) burden of variants within a single gene per cohort, and (4) unique combinations of variants across two or more genes per cohort.\n\nResultsWith few exceptions when using the relaxed filter, and which are likely the result of confounding by population stratification, NS patients did not have significantly increased burden of variants in Mendelian NS genes in comparison to a reference cohort, nor was there any evidence of oligogenicity. This was true when using both the relaxed and stringent variant pathogenicity filter.\n\nConclusionIn our study, the burden or particular combinations of low-frequency or rare protein altering variants in previously implicated Mendelian NS genes cohort does not significantly differ between North American patients with NS and a reference population. Studies in larger independent cohorts or meta-analyses are needed to assess generalizability of our discoveries and also address whether there is in fact small but significant enrichment of risk alleles or oligogenicity in NS cases undetectable with this current sample size. It is still possible that rare protein altering variants in these genes, insufficient to cause Mendelian disease, still contribute to NS as risk alleles and/or via oligogenicity. However, we suggest that more accurate bioinformatic analyses and the incorporation of functional assays would be necessary to identify bona fide instances of this form of genetic architecture as a contributor to the heritability of NS.