Fig. 3: Pathogenic variants by clinical criteria.

A Stacked bar plot indicating the fraction of individuals with a pathogenic variant identified and split by the nine clinical filtering criteria. As some individuals fulfilled multiple criteria we split them by diagnosis. This resulted in 292 total combinations of individuals and clinical criteria and 39 such combinations for individuals with a genetic diagnosis. To test whether certain criteria are enriched for genetic findings, we calculated p values assuming an equal diagnostic rate of 39/292 (red dotted line) in a simple Bernoulli experiment using a binomial test. Categories “hereditary disorders” (p ~ 0.000014) and “gout” (p ~ 0.023) showed nominally significant enrichment. The “hereditary disorders” category remained significant after adjusting for multiple testing. B Waffle plot comparing the nine filtering criteria and the gene in which a variant has been identified. Note that the 39 combinations of individuals and criteria are now also split by gene, because two individuals in the cohort had multiple pathogenic variants, resulting in 41 combinations. The two significant categories from A are all explained through variants in the COL4A3, COL4A4, and COL4A5 genes. Interestingly all three UMOD variants identified fall in the “interstitial nephritis” category, with one of them additionally classified as “nephrosclerosis”. Variants affecting HNF1B are either dispersed through four categories with none of them in the hereditary category, confirming both the variability in the HNF1B-associated disorders and the often sporadic nature of the CNVs (17q12 microdeletion/-duplication syndromes). Compare File S3 [21] for full variant details. C Violin and scatter plots comparing the kidney function parameters from Fig. 1D between individuals with a genetic variant identified (34) or not (237; green circles). Individuals with a COL4A3, COL4A4, or COL4A5 variant are presented in red and with variants in other genes in blue. Individuals with two variants are marked as diamonds. Individuals with IgA nephropathy are marked with yellow margin (compare also Fig. S2). The ACR at GCKD study inclusion is significantly higher in individuals with a genetic variant identified (two-sided Wilcoxon signed-rank test). D Upset plot showing the overlaps for individuals with a suspected “hereditary diagnosis” (as used for filtering), our finding of a pathogenic variant (“diagnosis”) and kidney biopsies performed. Overall in only six individuals with a confirmed diagnosis a kidney biopsy had been performed previously which likely raised the suspicion of an underlying genetic disorder. In 12 individuals with kidney biopsy where we identified pathogenic variants no suspicion of a hereditary disease was issued.