Article https://doi.org/10.1038/s41467-023-37985-w Multi-population genome-wide association study implicates immune and non-immune factors in pediatric steroid-sensitive nephrotic syndrome Received: 15 September 2022 A list of authors and their affiliations appears at the end of the paper Accepted: 10 April 2023 1234567890():,; Pediatric steroid-sensitive nephrotic syndrome (pSSNS) is the most common 1234567890():,; childhood glomerular disease. Previous genome-wide association studies Check for updates (GWAS) identified a risk locus in the HLA Class II region and three additional independent risk loci. But the genetic architecture of pSSNS, and its genetically driven pathobiology, is largely unknown. Here, we conduct a multi-population GWAS meta-analysis in 38,463 participants (2440 cases). We then conduct conditional analyses and population specific GWAS. We discover twelve sig- nificant associations—eight from the multi-population meta-analysis (four novel), two from the multi-population conditional analysis (one novel), and two additional novel loci from the European meta-analysis. Fine-mapping implicates specific amino acid haplotypes in HLA-DQA1 and HLA-DQB1 driving the HLA Class II risk locus. Non-HLA loci colocalize with eQTLs of monocytes and numerous T-cell subsets in independent datasets. Colocalization with kidney eQTLs is lacking but overlap with kidney cell open chromatin suggests an uncharacterized disease mechanism in kidney cells. A polygenic risk score (PRS) associates with earlier disease onset. Altogether, these discoveries expand our knowledge of pSSNS genetic architecture across populations and provide cell-specific insights into its molecular drivers. Evaluating these associations in additional cohorts will refine our understanding of population specificity, heterogeneity, and clinical and molecular associations. Pediatric steroid-sensitive nephrotic syndrome (pSSNS) is a rare immunosuppressants which allow some to achieve remission of their disease of the glomerular filtration barrier. Its incidence ranges proteinuria7, 10–17. from 1.15–16.9 cases in every 100,000 children, occurring most fre- Despite intensive investigation, there are no known monogenic quently in South Asian populations1. pSSNS causes massive protei- forms of pSSNS to illuminate its pathobiology. However, we know that nuria and increases the risk of thromboembolism, sepsis, and immune dysregulation is a major contributor18,19. But determining progression to chronic kidney disease (CKD)/end-stage kidney dis- causal immune factors via case-control studies of cytokines profiles, ease (ESKD)2–7. And those progressing to ESKD have increased odds cell types, and transcriptomic signatures is challenging. The dynamic of recurrent NS in their allograft8. pSSNS is impactful across the responses of the immune system at different disease stages and to lifespan—31–50% of those affected have relapses in adulthood9. Much various stimuli make it difficult to determine whether observed dif- of pSSNS’s morbidity is related to side effects of the non-specific ferences are causal, correlated, or due to independent biological/ e-mail: matthew.sampson@childrens.harvard.edu Nature Communications | (2023)14:2481 1 Article https://doi.org/10.1038/s41467-023-37985-w The fixed-effect meta-analysis was performed with METAL. Its P value is from a two-sided inverse-variance weighted meta-analysis and the Het χ2 P value is from a two-sided chi-square test for heterogeneity. The meta-regression was performed with MR-MEGA, and environmental factors. And kidney tissue in children is rarely available analysis was performed on all loci from the discover meta-analysis with exception of rs412175 and rs2637678. rs56117924 and rs2637681 were used in the conditional analysis, respectively. rs2256318 and rs1794497 are ~1.3 Mb apart with r2 < 0.13 across all the 1000 heterogeneity in allelic effects that are correlated with GWAS populations and is from a test of deviance of the full model compared to the model excluding population principal components. The residual heterogeneity is the deviance of the full model. Conditional Genomes Project populations, and with an r2 = 0.04 when combining all the 1000 Genomes samples. MR-MEGA results are not available for conditional analysis. Novel loci *, EA effect allele, NEA non-effect allele, OR [95% CI] = Odds ratio with 95% confidence Residual Het P value all tests are two-sided and approximated by a chi-square distribution. The MR-MEGA P value tests the association of the SNP and is from test of deviance of full meta-regression model compared to the null model. Its population Het P value measures the to determine intrarenal, molecular drivers of pSSNS. Previous GWAS have discovered four pSSNS risk loci20–24. In each GWAS, the top risk locus is in the HLA Class II region. Two other loci are plausibly immune-related, with the closest genes being Calcium 5.24 × 10−3 Homeostasis Modulator Family Member 6 (CALHM6)25 and TNF 0.30 0.43 0.07 0.59 0.82 0.87 0.18 Superfamily Member 15 (TNFSF15)26. The lead SNP of the fourth locus is '- - within nephrin (NPHS1), a fundamental glomerular gene implicated in Mendelian NS27. These studies are illuminating but limited by smaller sample sizes, primarily population-specific analyses, and limited post- Population Het P value GWAS analysis. Here we conducted a large and diverse GWAS of pSSNS to discover and more fully characterize disease-associated genetic variation and unravel its pathogenesis at the interface of the immune system and kidney. 4.99 × 10−4 2.43 × 10−4 7.76 × 10−4 0.34 0.26 0.24 0.74 0.16 Results - - We conducted a multi-population, fixed-effect, inverse-variance, meta- Meta-regression analysis across twelve GWAS datasets comprised of 2440 cases and 36,023 controls of Admixed American, African, East Asian, European, 2.30 × 10−24 6.85 × 10−82 −12 1.08 × 10−09 2.06 × 10−12 4.27 × 10−10 5.57 × 10−12 1.61 × 10−10 4.90 × 10 Maghrebian, and South Asian populations (Fig. 1, Supp. Fig. 1, Supp. P value Table 1).To account for population-driven effect heterogeneity, we also performed a meta-regression with MR-MEGA28. Given the - - increased power in the presence of heterogeneity across populations, we identified significant loci using MR-MEGA results. Eight loci (four Het χ2 P value new, and all outside HLA) were significant (MR-MEGA p < 5 × 10−8) 7.30 × 10−5 4.18 × 10−4 (Table 1, Fig. 2A, Supp. Fig. 2). The lead SNPs of the novel loci were all intronic: (1) rs7759971 in Abelson Helper Integration Site 0.09 0.04 0.03 0.02 0.33 0.62 0.52 0.18 1(AHI1; p = 4.90 × 10−12); (2) rs55730955 in CD28 molecule (CD28; p = 4.27 × 10−10); (3) rs8062322 in C-type Lectin Domain Containing 16 A (CLEC16A; p = 1.61 × 10−10); (4) rs28862935 in betacellulin (BTC; 6.79 × 10−52 1.00 × 10−10 2.92 × 10−10 3.59 × 10−11 −22 1.48 × 10−10 4.51 × 10−81 2.78 × 10−11 −13 9.71 × 10−18 p = 1.08 × 10−9). The remaining three significant loci located outside of 4.15 × 10 1.28 × 10 P value the HLA region were previously reported23,24. The associations found Fixed-effects meta-analysis near NPHS1 and TNFSF15 are driven by overlapping samples from Jia Table 1 | Genome-wide significant SNPs from multi-population meta-analysis 0.46 [0.42, 0.50] 0.79 [0.73, 0.85] 0.79 [0.73, 0.85] 0.74 [0.68, 0.81] 0.75 [0.68, 0.81] 2.04 [1.89, 2.27] 1.49 [1.36, 1.64] 1.65 [1.49, 1.83] 1.34 [1.24, 1.44] 1.41 [1.27, 1.56] OR [95% CI] NEA G G G C C A T T T T EA G C C C A A A A T T Position (hg19) 2:204585956 6:135746884 19:36342103 6:116787378 4:75693465 9:117579457 6:32649180 16:11092319 6:32627714 6:31381519 Fig. 1 | Flowchart of study design. 12 datasets across six populations were used for population-specific and multi-population GWAS meta-analyses. The population assignment and number of cases for each dataset are indicated (yellow=European rs55730955 rs28862935 rs2637678c rs10817678 rs1794497d rs2256318d rs8062322 rs1063355 rs7759971 (EUR), green=African (AFR), blue=East Asian (EAS), orange=South Asian (SAS), b Top SNP rs412175 purple=Maghrebian (MAG), red=Admixed American (AMR)). Post-GWAS analyses Conditional meta-analysis Discovery meta-analysis include colocalization with both kidney and immune eQTL datasets and overlap of SNPs within credible sets with single-cell kidney and immune open chromatin (ATAC-seq). HLA imputation with HLA-TAPAS was used to identify classical alleles and specific amino acids associated with pSSNS, followed by modeling of the HLA Nearest gene protein and stability predictions. Dataset summary statistics were used to generate HLA-DQB1 HLA-DQB1 CLEC16A* CALHM6 TNFSF15 polygenic risk scores using PRS-CSx and associations with clinical covariates were NPHS1 CD28* MICA* interval. AHI1* BTC* tested. pSSNS= pediatric steroid-sensitive nephrotic syndrome, eQTL = expression quantitative trait loci. Nature Communications | (2023)14:2481 2 Article https://doi.org/10.1038/s41467-023-37985-w A) 80 HLA-DQB1 6.85 x 10-82 -log10(P) 60 40 CALHM6 2.06 x 10-12 NPHS1 4.99 x 10-23 AHI1 20 4.90 x 10-12 TNFSF15 CLEC16A CD28 BTC 5.57 x 10-12 4.27 x 10-10 1.08 x 10-9 1.61 x 10-10 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 16 18 20 22 B) 50 HLA-DQB1 C) 50 6.79 x 10-52 40 40 30 30 -log10(P) -log10(P) AHI1+ 20 MICA 20 1.63 x 10-10 9.71 x 10-18 ELMO1+ MORF4L1 1.72 x 10-8 3.88 x 10-8 10 10 0 0 1 2 3 4 5 6 7 8 9 10 12 14 16 18 22 1 2 3 4 5 6 7 8 9 10 12 14 16 18 22 D) rs55730955 (CD28) rs28862935 (BTC) rs7759971 (AHI1) rs8062322 (CLEC16A) rs2256318 (MICA) P=2.33x10-3 P=0.03 Admixed American African P=7.36x10-10 P=2.03x10-4 P=2.15x10-5 P=2.82x10-8 P=2.20x10-16 East Asian P=4.12x10-5 P=3.89x10-6 P=0.01 P=1.04x10-3 European P=0.01 P=2.72x10-3 P=0.03 Maghrebian P=0.01 P=1.15x10-4 P=0.02 P=5.74x10-4 South Asian P=3.59x10-11 P=1.00x10-10 P=1.28x10-13 P=2.78x10-11 P=9.71x10-18 Meta-Analysis 0.5 1 1.5 1 2 3 1 2 3 0.4 0.6 0.8 1 1 1.5 2 OR (95% CI) Fig. 2 | GWAS results. All loci are labeled by nearest gene with novel associations in single-population odds ratios with 95% confidence interval for novel multi- red. A Multi-population meta-analysis of 2440 cases vs. 36,023 controls. The P value population significant SNPs. The P value for MICA is from the conditional analysis from test of deviance of full meta-regression model compared to the null model with COJO, Maghrebian, and Admixed American P values are from logistic regres- using MR-MEGA. B Multi-population conditional meta-analysis. The P value from sion, and the rest are from inverse-variance fixed-effects meta-analysis with METAL. multiple linear regression with COJO. C European meta-analysis of 674 cases vs. All P values in A–D are unadjusted for multiple testing and all tests are two-sided. 6817 controls. Discoveries that included the summary statistics from suggestive Number of cases in each analysis: Admixed American n = 98, African n = 109, East SNPs available from Dufek et al. are indicated with + and only novel associations are Asian n = 1311, European n = 674, Maghrebian n = 55, South Asian n = 193, Meta- labeled. The P values are from meta-analysis with METAL. D Multi-population and Analysis n = 2440. et al; however, this is an independent replication of the CALHM6 locus. associated with an increased risk in Europeans and those at the After conditioning on the lead SNPs, two more significant loci TNFSF15 and NPHS1 loci were associated with an increased risk in East emerged: (5) rs1794497 upstream of HLA-DQB1, (p = 6.79 × 10−52); (6) Asians (Supp. Fig. 5, Supp. Fig. 6). The remaining loci showed similar rs2256318 in an intron of MHC Class I Chain-related Gene A (MICA; effects across populations (Fig. 2D). Finally, there were 20 novel sug- p = 9.70 × 10−18) (Fig. 2B, Supp. Fig. 3). Population-specific GWAS meta- gestive loci (MR-MEGA p < 1 × 10−5) in the multi-population GWAS analysis discovered two additional significant loci in Europeans (Supp. Table 4, Supp. Table 5). On a liability scale and excluding HLA, (Fig. 2C, Supp. Table 2-3, Supp. Fig. 4): The lead SNPs were in introns of European heritability was 0.04 [CI: −0.08, 0.16] and East Asian herit- (7) rs111796602 in an intron of Engulfment and Cell Motility 1 (ELMO1; ability was 0.12 [CI: 0.04, 0.21], with large confidence intervals likely p = 1.72 × 10−8) and (8) rs12911841 in an intron of Mortality Factor 4 Like due to small effective sample sizes. 1 (MORF4L1; p = 3.88 × 10−8). Loci with population-driven heterogeneity A number of insights emerged from evaluating disease asso- were observed at three loci. Variants at the CALHM6 locus were ciations, functions, and expression patterns of the lead SNPs and/or Nature Communications | (2023)14:2481 3 Article https://doi.org/10.1038/s41467-023-37985-w the closest genes at the novel non-HLA loci. First, PheWAS significantly expressed in the human podocytes (https://atlas.kpmp. using Open Target Genetics (http://genetics.opentargets.org)29 org). CLEC16A is also involved in autophagy, mitophagy, and endoly- found that SNPs at most loci were associated with diverse white sosomal trafficking in multiple cell types39,40. Furthermore, it is blood cell traits, atopic disorders, and autoimmune conditions. also in close proximity to CIITA, a master transcription factor of HLA For example, among the strongest associations with the lead SNPs class II genes41 and Dexamethasone Inducible Transcript (DEXI), a at the following loci include: CLEC16A, CD28, MICA, and ELMO1 with glucocorticoid-induced gene42. eosinophil counts; AHI1 with monocyte and neutrophil counts, We next turned to discovering specific variants and genes driving asthma, and hay fever (also shared by CD28); and MICA with type 1 these GWAS loci and discerning whether they are acting in immune diabetes. cells, kidney cells, or both. Second, while most of these genes are primarily known for their First, we conducted colocalization with eQTL data from two role in immunity, many also have known roles in kidney diseases and functionally distinct kidney compartments (glomerulus and tubu- cells. Common AHI1 variants are associated with atopy, lupus, and lointerstitium; NEPTUNE43), multiple tissues from GTEx44, and immune diverse immune cell traits29. But rare, pathogenic AHI1 coding variants cells of healthy adults from DICE45 and BLUEPRINT46. Overall, pSSNS cause the monogenic ciliopathy Joubert Syndrome, which includes GWAS SNPs demonstrated significant enrichment in multiple immune cystic kidney disease30. ELMO1 participates in Rac1 pathway activation cell eQTLs, led by a 69× and 62× increased odds of being monocyte and actin cytoskeletal rearrangement31, is expressed in podocytes32, and CD4+memory Treg eQTLs, respectively (Fig. 3). On an individual and is associated with diabetic nephropathy33. CD28, a T-cell glyco- gene level, seven genes colocalized with immune cell eQTLs (Fig. 3, protein, binds a co-stimulatory molecule B7-1 (CD80) on antigen- Supp. Table 6). Three genes were closest to the lead GWAS SNP at their presenting cells. B7-1 is expressed on human podocytes in some forms respective locus—CALHM6, AHI1, and TNFSF15. Each were significantly of nephrotic syndrome and blocking the B7-1/CD28 interaction with a colocalized with monocyte eQTLs. AHI1 also colocalized with many CTLA-4 immunoglobulin can ameliorate proteinuria34. MICA is T-cell subsets and naive B cells. Finally, a suggestive locus on chro- expressed in kidney endothelium, binds and activates cytotoxic mosome 17 colocalized in CD4+memory Treg cells with two distinct CD8+T cells and NK cells, and has increased glomerular expression in genes—Gasdermin B (GSDMB) and ORMDL sphingolipid biosynthesis lupus35. BTC contributes to inflammation by binding to epidermal regulator 3 (ORMDL3). The GSDMB/ORMDL3 locus is associated with growth factor receptor36, a gene whose kidney expression is upregu- multiple autoimmune disorders and eosinophilic inflammation-driven lated following kidney injury37. CLEC16A takes part in the B cell asthma47. In asthma, higher GSDMB expression is correlated with receptor-dependent HLA-II pathway in human B cells38 but is also increased interferon signaling and MHC class I antigen presentation48. Colocalized Genes eQTL Enrichment Estimate Monocyte BP Neutrophil T cell Monocytes, classical T cell, CD4+, memory Treg T cell, CD4+, naive T cell, CD4+, TFH T cell, CD4+, naive Treg Monocytes, non-classical B cell, naive DICE T cell, CD4+, Th17 T cell, CD8+, naive T cell, CD4+, Th1/17 T cell, CD4+, Th2 T cell, CD4+, Th1 T cell, CD4+, naive [activated] NK cell, CD56dim CD16+ T cell, CD8+, naive [activated] Whole Blood GTEx Artery Aorta Skin Not Sun Exposed Adipose Visceral Omentum NEP Tubulointerstitium Glomerulus Ln 00 A1 M 1 RM -2 FS L3 L I -D B H L SD 6 5 LH I1 0 1 2 3 4 c- 2 7 G M F1 A M CA AH O YB TN D NC Q 0 0.5 1 RCP log(odds ratio) Fig. 3 | Colocalization of SSNS GWAS and eQTL datasets. Each eQTL dataset is estimates from fastENLOC are based on genome-wide summary statistics from labeled with colocalized loci (left) and enrichment estimates (right). The source of GWAS and include a shrinkage parameter that results in 0 enrichment for multiple each eQTL dataset is labeled on vertical gray bars. BP BLUEPRINT, NEP NEPTUNE. tissues/cell types. Estimates are presented as the logarithm of the odds ratio ± Genes with regional colocalization probability (RCP) > 0.2 in at least one tissue/cell standard error. logOR = 2 ~ OR = 7.5, logOR=3 ~OR = 20.1, logOR = 4 ~ OR = 54.6. are included. pSSNS GWAS loci that colocalized with tissue/cell-type eQTLs are eQTL sample sizes: NEPTUNE glomerulus n = 240, tubulointerstitial n = 311, BLUE- indicated by black dots, with larger dots indicating higher RCP. GTEx tissues PRINT n = 200 DICE n = 91. without associations are excluded from this figure (see Supp. Fig. 7). Enrichment Nature Communications | (2023)14:2481 4 Article https://doi.org/10.1038/s41467-023-37985-w A) Amino Acid Haplotype Frequency Predicted DQA1_47 DQA1_52 Case Control OR (95% CI) P-value Stability Change lysine histidine 0.28 0.12 3.56 [3.12, 4.06] 1.09 x 10 -78 -3.64 cystine arginine 0.34 0.31 1.93 [1.71, 2.18] 8.77 x 10 -27 -2.67 glutamine arginine 0.13 0.18 1.69 [1.45, 1.97] 2.90 x 10 -11 -0.05 arginine serine 0.25 0.39 Reference - - B) Fig. 4 | HLA-DQA1 amino-acid associations and stability prediction. A Increased predicted stability change indicate decreasing stability. B Protein structure for the risk and predicted stability change of the two-amino-acid residue haplotypes at reference haplotype arginine47-serine52 (left, blue) and lysine47-histidine52 (right, HLA-DQA1 positions 47 and 52. Odds ratios and P values (two-sided) are from a joint red). The residues in green display a potential interacting amino acid with mutated logistic regression with arginine47-serine52, the most common, set as reference, amino acids. The color scheme for interactions (dashed lines) is as follows: cyan for adjusting for population-specific principal components and continental popula- Van der Waals [VDW], red for hydrogen bonds, green for hydrophobic bonds, sky tions. The reference haplotype confers the strongest protection (i.e., odds ratios blue for carbonyl bonds, and orange for polar bonds. Amino acids displayed with indicate increase in risk compared to arginine47-serine52). Decreasing values of the no visible bonds indicate a prediction of weak VDW bonds. Notably, there was no colocalization with kidney eQTLs despite suffi- increase in odds of pSSNS (Fig. 4A). HLA-DQA1 position 47 is located on cient sample sizes (Supp. Fig. 7). the outside of the peptide-binding groove and acts as a regulator of We then created a 95% credible set for all non-HLA significant loci binding stability, which, when altered, has been suggested to mediate and assessed their overlap with ATAC-seq derived open chromatin the development of autoimmune disorders53. Arginine at HLA-DQA1 data from immune49 and kidney cells50, 51 (Supp. Table 7). The SNPs with position 52 has been associated with autoimmune disorders, including the highest posterior inclusion probability (PIP) for AHI1, rs7759971 type 1 diabetes54. (PIP = 0.40), overlapped with open chromatin of multiple immune cell We then used DynaMut255 to model the impact of the types, including CD34 + cells, common lymphoid and myeloid pro- 47lysine−52histidine haplotypes on protein structural stability (Fig. 4B). genitors, hematopoietic stem cells, and multipotent progenitors. This is quantified by Delta Delta G (ddG), where ddG <0 predicts Similarly, the top SNP for CD28, rs55730955 (PIP = 0.68), overlapped unstable structure. The haplotype consisting of lysine (47) and histi- with CD4+and CD8+open chromatin. The top PIP SNP for BTC, dine (52) predicted the most instability (ddG = −3.64). Notably, the CLEC16A, and TNFSF15 had no overlap with open chromatin. However, predicted increase in protein instability and increased odds of disease each locus had individual SNPs with lower PIPs that overlapped with for each haplotype were concordant. This suggests pSSNS-associated both immune and kidney cell open chromatin. haplotypes increase the odds of disease by increasing the instability of We next fine-mapped the HLA risk locus to discover classical HLA HLA-DQA1 and altering its ability to properly form a stable HLA-II alleles and amino acids associated with pSSNS (Supp. Table 8). We first molecule. imputed across the extended MHC region using a multi-population Finally, we generated a multi-population pSSNS polygenic risk HLA imputation panel52, resulting in 640 classical HLA alleles, 4513 score (PRS) using summary statistics of 1974 cases and 20,039 controls amino acids in HLA proteins, and 49,321 SNPs in the extended MHC from European, East Asian, African, and South Asian populations. We region for association. We used population-specific and multi- tested the association of the PRS with demographic and clinical phe- population SNP-level logistic regression, to identify specific SNPs notypes in 233 European children with sufficient clinical data from the and classical alleles associated with pSSNS (Supp. Table 9, Supp. EU-European sub-cohort, adjusting for four genetic principal compo- Note 1, Supp. Table 10, Supp. Fig. 8). nents. The highest PRS quartile had significantly lower age of onset We next turned to discovering specific HLA amino-acid positions (Q4: 4.9 years) compared to the lowest quartile (Q1: 6.9 years, most associated with risk of pSSNS through logistic regression analysis p = 2.79 × 10−3, Supp. Table 12). We did not find a significant association of all residues at each position. Amino-acid position 47 in HLA-DQA1 between PRS and sex or relapse pattern. Of note, we found concordant was most strongly associated with pSSNS (Pomnibus = 7.73 × 10−83) results using a PRS generated using discovery GWAS from a PRS score (Supp. Table 11, Supp. Fig. 9). Arginine was the most frequent amino- generated from European GWAS only (Supp. Table 12, Methods). acid; a substitution to lysine conferred the greatest disease risk (p = 5.70 × 10−80; OR [95% CI] = 3.62 [3.17—4.14]). A second association Discussion in near-perfect linkage disequilibrium was identified at HLA-DQA1 A number of important discoveries emerged from this study. First, we position 52 (p = 1.14 × 10−82). Arginine was again the most common identified seven novel pSSNS loci—four novel loci from our multi- amino-acid at this position, and a substitution to serine conferred the population meta-analysis (BTC, AHI1, CD28, and CLEC16), one locus greatest protection from risk (p = 1.00 × 10−28; OR = 0.53 [0.47—0.59]). from the corresponding conditional analysis (MICA), and two addi- After conditioning, an independent association was discovered at HLA- tional novel loci unique to the European meta-analysis (ELMO1, DQB1 position 26 (p = 3.22 × 10−13). A change from the most common MORFL1). Second, we found that while the immunological connections amino-acid leucine to glycine conferred the most significant protec- with the lead SNPs and closest genes in these newly discovered loci are tion (p = 4.75 × 10−12; OR = 0.64 [0.60—0.73]). A haplotype analysis well-established, most of them also have a bona fide, but overall less identified the 47lysine−52histidine haplotype was associated with greatest understood role, in kidney cells and diseases. Identifying the genes, Nature Communications | (2023)14:2481 5 Article https://doi.org/10.1038/s41467-023-37985-w cells, and organ systems with which each of these identified risk loci act region has also been implicated in association with eGFR, a common will be an important future step. complex kidney trait28,58. How is altered expression of these genes For instance, is the lead SNP near AHI1 in fact altering the function involved in both a rare glomerular disease and a common, complex of AHI1 itself? And if so, how will we come to understand how rare, kidney trait? We don’t currently know. Ultimately, applying in silico coding variants in this gene cause a structural, cystic kidney disease methods to high-quality, human-derived immune- and kidney-omics while a common, non-coding variant impacting the same gene con- datasets should help pare down candidate alleles, genes, cell types, tributes to an immunologically-mediated, acquired condition of the and mechanisms to a manageable number for subsequent experi- kidney? The availability of single-cell omics data from larger samples mental studies in cells and model systems. sizes and pediatric kidney tissue will be critical to help sort this out. Finally, the association of higher PRS with younger age of onset These new datasets will also help post-GWAS studies like colocalization suggests that a stronger genetic predisposition to disease lowers the with eQTL and open chromatin, as we hypothesize that the paucity of threshold of an individual to develop pSSNS in the context of envir- kidney eQTLs we observed may be due to mapping pSSNS GWAS data onmental factors and may ultimately help share clinical screening and to molecular datasets that do not adequately represent rare kidney cell care. We must now evaluate this PRS in other cohorts, such as in types or changes that occur in the childhood age. In another example, cohorts of pSSNS from other global populations, adults with the dis- we have now identified a two amino-acid haplotype in HLA-DQA1 that ease, or children with other forms of childhood-onset NS. Studying the increases risk of pSSNS. By what mechanism is it doing this? molecular correlates of a high PRS score could also shed light on Third, while our field has been long-focused on the role of T cells pathobiology driven by an increased burden of genetic risk of this in pSNNS and more recently B cells, our results now suggest that disease. monocyte and eosinophil gene dysregulation may also be a potential There are limitations to this study. Cases and the reference contributor to the pathogenesis of pSSNS. Are we observing a patho- population were not all genotyped on the same SNP array. While we logic signature from resident monocytes and/or circulating mono- used a robust strategy to account for the use of different SNPs arrays cytes. Alternatively, are some single-cell immune eQTLs a proxy for containing different SNPs, this still adds heterogeneity. Heterogeneity certain kidney subtypes that we don’t observe in our current bulk-level is also added by the age of onset pSSNS not being identical for all analysis? While SSNS is sometimes observed in the hypereosinophilic groups, with the Columbia-originating cohort including patients less condition Kimura disease56, its etiology unknown. Defining the than 21 as opposed to 18 for the other groups. And for the most part, mechanism by which genetically driven changes in these cell types we did not recruit healthy controls. Rather, we relied on available contribute to pSSNS onset is an important area of future inquiry. reference populations, assuming that, as a rare disease, pSSNS cases Our analysis revealed a lack of significant colocalization between were absent within them. Each of these factors would be predicted to genetic variants associated with pSSNS and those associated with gene reduce power for discovery. Our ability to accurately measure herit- expression in kidney tissues. This could be consistent with the ability is limited. This is because sparse signals with strong effects can understanding that pSSNS is primarily a disease of immune dysregu- lead to less efficient estimates and low sample sizes can downwardly lation, with the kidney being the end organ affected. It is thus plausible bias the results59. Moreover, heritability can vary among populations. that most of the genetic risk for pSSNS is mediated through genes and Thus, to better understand heritability of pSSNS within and across pathways in immune cells. This is supported by previous research that populations, it is crucial to increase sample sizes across all popula- has shown that transplanting healthy kidneys into patients with a his- tions. Finally, the sample sizes for South Asian, Maghrebian, and Afri- tory of NS can result in recurrence of disease. Furthermore, a case can could result in overfitting of summary statistics and were too low report described that retransplanting the now-diseased kidney into a to allow us to perform rigorous post-GWAS analysis of their results on patient without NS resulted in disease cessation in the transplanted a population-specific basis. Our novel genome-wide significant loci kidney, suggesting that an abnormal immune system alone is sufficient revealed consistent directionality across all populations, but with to cause pSSNS8,57. varying magnitudes and significance. Obtaining independent datasets Nevertheless, we posit that, as exemplified by the NPHS1 risk with sufficient sample sizes from each population will enhance our locus, there may be genetic factors that contribute to pSSNS through understanding of population heterogeneity and refine estimates of regulatory effects on kidney cells. Larger GWAS and eQTL studies will effect size. provide greater statistical power, which may uncover colocalized In conclusion, the discoveries emerging from our global GWAS of variants in the kidney. In addition, eQTL analysis using single-cell RNA- pSSNS expand our knowledge of the genetic architecture of this dis- seq should provide more power to detect eQTLs in rare cell types, such ease and accelerate our understanding of its molecular underpinnings as podocytes and glomerular endothelium. These signals that may be and clinical implications. currently missed with the bulk methods. We may also more readily detect kidney cell impact from these pediatric SSNS GWAS data if use Methods kidney cell omics datasets derived solely from children. Integrating This research was conducted with the informed consent of all study pediatric-derived GWAS and molecular data (e.g. eQTL, open chro- participants and had ethical approval from the Boston Children’s matin) derived by patients of the same age may capture state- Hospital IRB. dependent relationships that are obscured when using adult omics Figures were generated with R (v3.6.3) and ggplot2 (v3.3.5). data with pediatric GWAS summary statistics. Finally, it is important to note that regulation of genes in kidney cells by risk SNPs may occur GWAS data summary through mechanisms that are not linked to levels of gene expression, Recruitment of samples and statistical analyses varied by recruiting such as allele-specific expression of NPHS123. Therefore, future studies group. Details for each dataset are described below and in Supp. should also consider looking for evidence of allele-specific expression, Table 1. splice QTLs, protein QTLs, or other forms of dysregulation to uncover the impact of GWAS alleles on kidney cells. GWAS data from NEPHROVIR/EU. Sample collection and genotype Fourth, we discovered specific amino-acid changes in HLA-DQA1 calling were done at Sorbonne Université in Paris. Pediatric steroid- and -DQB1 associated with pSSNS that should empower subsequent sensitive nephrotic syndrome was defined as proteinuria >0.25 g/mmol, studies to illuminate pathomechanisms at the risk locus that has been serum albumin <25 g/L (< 30 in France), full response within four weeks identified in every GWAS of pSSNS to date. But change in HLA-DQA1 of 60 mg/m2/day of oral prednisone or prednisolone, and age of onset and HLA-DQB1 gene expression due to genetic variation in the MHC <18 years old. 244 previously reported European patients from the Nature Communications | (2023)14:2481 6 Article https://doi.org/10.1038/s41467-023-37985-w NEPHROVIR study21 were combined with 159 newly recruited European batch effects62. Pre-imputation QC was conducted using McCarthy patients recruited from France, Lithuania, Poland, Russia, Italy. Healthy Tools v4.3 with the TOPMed reference panel to check strand alignment adult controls (n = 300) were recruited from Lyon, France, and com- and allele assignment. Insertions and deletions were excluded prior to bined with population-matched controls from the 1000 Genomes imputation. Each population was imputed separately with cases and Project Phase 3 release (n = 493)60 and the 3Cites Cohort (n = 2000). controls were imputed together on the TOPMed Imputation Server There were also 56 sub-Saharan African cases with 451 African controls with the TOPMed r2 reference panel63–65. The QC was repeated after from the 1000 Genomes Project and 85 Maghrebian cases with 261 imputation and SNPs with low imputation quality (rsq <0.3) were Moroccan population-matched controls. Both were reanalyzed from a excluded. After imputation, UCSC Liftover66 was used to convert SNP previous report21. There were 160 Indian cases with 93 population- positions from each population dataset to build GRCh37 to match matched controls. Samples were genotyped on the Illumina Human the build of summary statistics from other analyses. The association OmniExpress or Illumina Omni 2.5 arrays. tests were completed using PLINK v1.9 under an additive model with principal component adjustment to account for population GWAS data from Columbia University (US Cohorts). Sample collec- stratification. tion and genotype calling was performed at Columbia University in New York. Cases were defined by local recruitment centers across the GWAS data from Columbia University (US Cohorts): US-European, US, Europe, and Brazil as either minimal change disease or non- US-African, US-South Asian, US-East Asian, and Admixed American. biopsied SSNS with age of disease onset <21. Five cohorts from Population was assigned by KING67 kinship analysis software and based Columbia University consisted of patients from European (ncases=371, on continental population as defined by the 1000 Genomes Project for ncontrols = 4359), East Asian (ncases = 17, ncontrols = 443), sub-Saharan all cases and controls68,69. Within each continental population (EUR, African (ncases = 65, ncontrols = 7344), South Asian (ncases =39, AFR, AMR, SAS, and EAS), we removed variants with genotype rate ncontrols = 534) and Admixed American (ncases = 109, ncontrols = 13,266) <99%, MAF < 0.01, and HWE P < 1 × 10−5. Each population was imputed populations. As defined by the 1000 Genomes Project, Admixed separately with the TOPMed r2 panel63–65. After imputation, we Americans (AMR) include the following populations: Puerto Rican in removed first-degree relatives using KING, and variants with R2 < 0.8, Puerto Rico (PUR), Columbian in Medellin (CLM), Peruvian in Lima MAF < 0.01, and HWE P < 1 × 10−5. Principal components were calcu- (PEL), Mexican Ancestry in Los Angeles (MXL). The genotyping of the lated with FlashPCA70. For cohorts with large case/control imbalances cases used multiple versions of MEGA (Multi-Ethnic Global Array) chips (Admixed American and US-African), we used the SAIGE logistic mixed that includes MEGA 1.0, MEGA 1.1, and MEGAEX. The controls that were model69 for calculating p value and generating summary statistics. genotyped on MEGA 1.0 were downloaded from NCBI dbGAP (IDAT Association tests for European, South, and East Asian were completed files) from the PAGE consortium61. The differences between the chips using PLINK v1.9 under an additive model with principal component were corrected first by mapping all the SNPs to a common cluster file in adjustment to account for population stratification62. Genome Studio for individual MEGA platforms and then using Snpflip software. GWAS data from Kobe University: Japanese, Korean, and Thai. Quality control and analysis of the Japanese dataset are previously GWAS data from Kobe University. Pediatric steroid-sensitive described in Jia et al.23. Samples were filtered for call rate <97%, nephrotic syndrome cases were defined as urine protein to creati- ambiguous sex, and IBD > 0.1875. Variants were filtered for info score nine ratio ≥2.0, serum albumin ≤2.5 g/dl, and complete remission with >0.5, missing >3%, MAF > 0.5%, and HWE P ≥ 0.0001 in controls. SNPs 4–6 weeks after starting 60 mg/m12 oral prednisolone per day and age were imputed with a Japanese reference panel with IMPUTE4 (v2.3.1). of onset <18 years old. Three GWAS studies of SSNS in Japanese For the Thai dataset, SNPs with MAF < 0.005, call rate <97%, or HWE (ncases = 987, ncontrols = 3206), Korean (ncases = 243, ncontrols = 4041) and P < 1 × 10−5 were removed. Individuals with missing rate > 3%, IBD > Thai (ncases = 65, ncontrols = 94) population were completed at The 0.1875 and PCA outliers were removed. For the Korean dataset, SNPs University of Tokyo, Japan. The Japanese GWAS data have been pre- with MAF < 0.01, call rate <99%, or HWE P < 5×10−8 for cases and viously reported22,23. The Thai dataset was genotyped with the Axiom <1 × 10−5 in controls were removed. Individuals with missing rate >4% or array. The Korean data was genotyped with the Affymetrix Axiom array IBD (PI_HAT) > 0.2. No outliers were removed from PCA inspection. for cases and Illumina OmniQuad chip for controls. Both Thai and Korean genotypes were imputed with the 1000 Gen- omes reference panel using SHAPEIT71 and IMPUTE272 and SNPs were Dataset QC, imputation, and GWAS filtered for info score <0.9 and 0.8 in Thai and Korean, respectively. Quality control, imputation, and GWAS were conducted separately for Logistic regression was performed with Plink v1.9. Sex and the first four each study location and population. GC lambda (GCλ) was used to principal components were used in the Japanese cohort. No covariates assess inflation in all studies. The final case and control sizes and the were adjusted for the Thai and Korean datasets and p values were number of variants tested can be found in Supp. Table 1 and Supp. adjusted for genomic control (GC). Fig. 1. Supplementary Fig. 11 shows matching of cases and controls in PCA plots. Manhattan plots and GC can be found in Supp. Fig. 12 and Population-specific and multi-population meta-analysis genome-wide significant hits resulting from dataset GWAS are in Supp. For each population-specific meta-analysis and the multi-population Table 13. meta-analysis, we conducted an inverse-variance, fixed-effect meta- analysis using METAL (v2011-03-25) with adjustment for population GWAS data from NEPHROVIR/EU: EU-European, EU-African, stratification (GC) on each input dataset and assessment for hetero- Maghrebian, Indian. Each file was quality controlled separately to geneity selected73. For within-population meta-analyses, we removed remove related individuals (IBD > 0.1875), low call rate (genotype rate variants with heterogeneity p value <0.05. All significant associations <98%), and cases with discordant sex. SNPs were quality controlled for were visually inspected and single SNPs that did not follow the allele frequency (MAF < 0.01), call rate (genotype rate <98%) in all expected LD trend and SNPs with within-population heterogeneity cohorts, and Hardy Weinberg equilibrium (HWE P < 1 × 10−5) in controls removed. only. The EU-European datasets were generated in multiple files and For the European meta-analysis, we included summary statistics were merged stepwise on the common subset of SNPs, with the pre- of only suggestive SNPs from a published GWAS in which the full data vious QC procedure reapplied after each merge. PCA plots were con- was not available24, increasing the European sample size to 1096 cases structed from PLINK v1.9 to identify population outliers and check for and 12,459 controls. Nature Communications | (2023)14:2481 7 Article https://doi.org/10.1038/s41467-023-37985-w Multi-population meta-regression with MR-MEGA For immune open chromatin, 76 samples from primary whole blood To account for and assess heterogeneous loci, we conducted a meta- were used resulting in 1000–100,000 FACS-purified cells (GSE74912). regression using MR-MEGA v0.228. We included three principal com- For kidney, kidney cortex from 5 patients undergoing nephrectomies, ponents, which captured the population structure across all twelve resulting in 35,286 cells (GSE151302). Open chromatin peaks were datasets. This allowed us to stratify heterogeneity into residual het- identified by the MACS2 (v2.2.7.1) peak calling algorithm and opti- erogeneity and heterogeneity that correlates with population. For each mized by gkmQC (v1.0)51. variant with heterogeneity that correlated with population, we visua- lized the dataset PCs from MR-MEGA with the dataset-specific log odds HLA imputation and analysis ratio from METAL. We adjusted for genomic control at the study level To fine-map the HLA region, we conducted HLA imputation with the and after meta-regression to account for population structure within four-digit multi-ethnic v2 reference panel on Michigan Imputation and between datasets. SNPs present in less than five studies were Server52. Cohorts were imputed individually to optimize population- excluded. GC lambda (GCλ) was used to assess inflation. Results tables specific structure within the HLA region. The imputed cohorts were include summaries from both METAL and MR-MEGA analyses (Table 1). then merged for multi-population associations. We used HLA-TAPAS All Manhattan plots were generated with the qqman R package v0.4.1 (v2020.05.02) ‘assoc’ module to conduct a logistic regression of the [doi: 10.21105/joss.00731.] and LocusZoom web tool74. All significant HLA region of the multi-population and population-specific datasets. loci are >1 Mb from each other with r2 < 0.1. Loci are labeled by For population-specific analyses, we adjusted for genotype-based nearest genes. principal components from Plink v1.962. The population-specific prin- cipal components and continental populations were included as cov- Conditional analyses ariates in the multi-population analysis. HLA-TAPAS was also used to To identify independent secondary significant loci at the candidate conduct a stepwise conditional analysis, conditioning on the locus loci, we used GCTA COJO (v.1.93.2beta)75,76 to conduct approximate with the smallest association p value. We additionally performed an conditional analyses based on cohort-specific meta-analysis summary omnibus test on the population-specific and multi-population cohorts statistics. Conditional analysis was conducted in each dataset, with an to assess significance by amino-acid position. LD reference generated from the dataset samples, due to differences in linkage disequilibrium structure between continental populations. HLA modeling Each cohort was conditioned for the eight independent loci identified To predict the reference (with arginine at position 47 and serine at from the initial meta-analysis. Multi-population meta-analysis of the position 52) structure of HLA-DQA1 we extracted the sequence of HLA- conditioned cohorts was repeated in METAL73 to assess multi- DQA1 from UNIPROT database (Uniprot ID: P01909). We used NCBI population genome-wide significant secondary loci after GCTA. BLAST against PDB database to find the closest structure associated with the amino-acid sequence of P01909. We identified the top hit as Heritability estimates 6PX6_A (HLA-TCR complex, E = 2x10−161) for the HLA-DQA1 sequence83. SNP-based heritability was estimated on a liability scale with LD score We extracted the PDB coordinates for chain A from the 6PX6 and regression (LDSC v1.0.1)77 using a population prevalence of 16/ visualized in PYMOL v2.5. Since the most common amino-acid haplo- 100,000 and excluding HLA [chr6:25,000,000-34,000,000]. We type in the control population was arginine (47) and serine (52), we used non-GC corrected population-specific meta-analysis summary performed mutagenesis using PYMOL to model the reference protein statistics from METAL and pre-computed LD scores generated 3D structure84. from the 1000 Genomes EUR or EAS samples. (https://alkesgroup. In brief, we used the mutagenesis tool from PYMOL and selected broadinstitute.org/LDSCORE/). the rotamer (most likely amino-acid conformation) for arginine and serine which showed the minimum number of clashes with nearby Colocalization of SSNS GWAS variants and eQTLs datasets atoms. Afterwards, we adjusted the conformation of nearby atoms We used fast enrichment estimation aided colocalization analysis (within 5 Angstrom) to minimum free state using ‘Clean’ command in (fastENLOC v1)78 for colocalization analysis with glomerular (n = 240) PyMOL which uses MMFF94 force field85. Though point mutations and tubulointerstitial (n = 311) eQTLs from nephrotic syndrome locally affect the conformation of the protein, they can result in tor- patients79, GTEx tissues (varied sample sizes), and immune eQTLs from sion, bending and stretching of the entire molecule. Therefore, we both BLUEPRINT46 (n = 200) and DICE45 (n = 91) databases. Posterior exported the protein structure to SPDBV software for further probabilities for SSNS GWAS variants were calculated from MR-MEGA refinement86. Z-scores using TORUS80. We used an LD panel from European and East We first fixed all the side chains of all amino acids to the best Asian 1000 Genomes samples to define haplotype blocks in the pSSNS rotamer conformation using the simulated annealing method. After- meta-analysis60,81. Enrichment of pSSNS GWAS variants in each eQTL wards, we performed energy minimization using GROMOS 96 force dataset was estimated using fastENLOC and subsequently informed field to extract the 3D coordinates that represent the lowest minimum prior probabilities for each analysis. For colocalization with our kidney energy conformation87. The refined protein structure of HLA-DQA1 was eQTLs, which had available raw data, we could identify multiple eQTLs then assessed for changes in stability of protein for both amino-acid per gene and multiple colocalized eQTLs at each locus. For all other combinations for each haplotype using “MULTIPLE MUTATION” in data, in which only summary statistics were available, we assumed at DynMut2 server55. The instability of HLA-DQA1 was evaluated using the most one colocalized SNP per loci. predicted ddG parameter which measures changes in Gibbs free energy between the folded and unfolded states and the change in Open chromatin annotation of credible sets folding when a mutation is present. The interaction among amino 95% credible sets were constructed for each independent locus iden- acids in reference and mutated structure were predicted using tified from the multi-population meta-regression with Bayes’ factors Arpeggio88 and visualized in PyMOL. reported by MR-MEGA. Posterior inclusion probability (PIP) was esti- mated by dividing each Bayes’ factor by the summation of Bayes’ fac- Polygenic risk score analysis tors across all variants within 1 Mb from the lead locus82. Construction of the multi-population PRS. To investigate genetic risk SNPs within 95% credible sets of our genome-wide significant loci across the genome, we generated a PRS using 1974 cases and 20,039 were evaluated for positional overlap based on the boundaries controls from the GWAS of European (US-European), East Asian of known open chromatin peaks in kidney50 and immune49 cell types. (Japanese, Korean, US-East Asian), African (US-AFR, EU-AFR), and Nature Communications | (2023)14:2481 8 Article https://doi.org/10.1038/s41467-023-37985-w South Asian (US-SAS, Indian) populations using PRS-CSx (v7-29- 4. Kerlin, B. A. et al. Epidemiology and risk factors for thromboembolic 2021)89. Populations with less than 100 cases (Maghrebian, Admixed complications of childhood nephrotic syndrome: a Midwest American) were excluded. The EU-EUR dataset was excluded from PRS Pediatric Nephrology Consortium (MWPNC) study. J. Pediatr. 155, calculations and was used as an independent test/train dataset, where 105–110 (2009). 80% of cases and controls were randomly selected for training and 20% 5. Hingorani, S. R., Weiss, N. S. & Watkins, S. L. Predictors of peritonitis for testing. While our goal of this analysis was to explore the rela- in children with nephrotic syndrome. Pediatr. Nephrol. Berl. Ger. 17, tionships between the PRS and clinical correlates within case cohorts, 678–682 (2002). not case/control prediction, we used prediction accuracy to optimize 6. Rheault, M. N. et al. AKI in children hospitalized with nephrotic the gamma-gamma priors and the global shrinkage parameter used in syndrome. Clin. J. Am. Soc. Nephrol. CJASN 10, 2110–2118 (2015). the PRS-CSx model. We varied the hyper parameters and chose the 7. Sato, M. et al. Prognosis and acute complications at the first onset of model with the best prediction accuracy (F-measure; Supp. Table 15). idiopathic nephrotic syndrome in children: a nationwide survey in The regression betas from the best model were used to weight the Japan (JP-SHINE study). Nephrol. Dial. Transplant. 36, population-specific PRS. 475–481 (2021). 8. Ding, W. Y. et al. Initial steroid sensitivity in children with steroid- Construction of the European PRS. We also generated a PRS from the resistant nephrotic syndrome predicts post-transplant recurrence. GWAS of European (US-European) populations using PRS-CS90. Similar J. Am. Soc. Nephrol. JASN 25, 1342–1348 (2014). to the multi-population method, the EU-EUR dataset was used as an 9. Korsgaard, T., Andersen, R. F., Joshi, S., Hagstrøm, S. & Rittig, S. independent test/train dataset, where 80% of cases and controls were Childhood onset steroid-sensitive nephrotic syndrome randomly selected for training and 20% for testing. We used prediction continues into adulthood. Pediatr. Nephrol. Berl. Ger. 34, accuracy to optimize the gamma-gamma priors and the global 641–648 (2019). shrinkage parameter used in the PRS-CS model. We varied the hyper- 10. Ishikura, K. et al. Morbidity in children with frequently relapsing parameters and chose the model with the best prediction accuracy (F- nephrosis: 10-year follow-up of a randomized controlled trial. measure; Supp. Table 14). Pediatr. Nephrol. Berl. Ger. 30, 459–468 (2015). 11. Fakhouri, F. et al. Steroid-sensitive nephrotic syndrome: from Clinical associations with PRS. The PRS was applied to pediatric childhood to adulthood. Am. J. Kidney Dis. J. Natl Kidney Found. 41, participants from the EU-European data for which clinical data were 550–557 (2003). available (n = 233). For each PRS (European and multi-population), we 12. Kyrieleis, H. A. C. et al. Long-term outcome of biopsy-proven, fre- split the samples into PRS quantiles. We tested significance of PRS quently relapsing minimal-change nephrotic syndrome in children. quartiles in the following models: Model 1: sex ~ PRS + age of onset + Clin. J. Am. Soc. Nephrol. CJASN 4, 1593–1600 (2009). relapse pattern + 4PCs (multiple logistic regression); Model 2: relapse 13. Skrzypczyk, P. et al. Long-term outcomes in idiopathic nephrotic pattern ~ PRS + age of onset + sex + 4PCs (multiple logistic regression); syndrome: from childhood to adulthood. Clin. Nephrol. 81, Model 3: age of onset ~ PRS + sex + relapse pattern + 4PCs (multiple 166–173 (2014). linear regression). The effect size, standard error, and p value for the 14. Trompeter, R. S., Lloyd, B. W., Hicks, J., White, R. H. & Cameron, J. S. PRS effect are reported. Long-term outcome for children with minimal-change nephrotic syndrome. Lancet Lond. Engl. 1, 368–370 (1985). Reporting summary 15. Aydin, M. et al. The long-term outcome of childhood nephrotic Further information on research design is available in the Nature syndrome in Germany: a cross-sectional study. Clin. Exp. Nephrol. Portfolio Reporting Summary linked to this article. 23, 676–688 (2019). 16. Lee, J. M., Kronbichler, A., Shin, J. I. & Oh, J. Review on long-term Data availability non-renal complications of childhood nephrotic syndrome. Acta The fixed-effects multi-population summary statistics (METAL) gen- Paediatr. Oslo Nor. 1992 109, 460–470 (2020). erated by this study have been deposited in the GWAS Catalog 17. Hjorten, R., Anwar, Z. & Reidy, K. J. Long-term outcomes of child- [GCST90258619]. The credible sets for significant variants and sig- hood onset nephrotic syndrome. Front. Pediatr. 4, 53 (2016). nificant colocalization results generated in this study are provided in 18. Shalhoub, R. J. Pathogenesis of lipoid nephrosis: a disorder of T-cell the Supplementary file. The raw GWAS data are protected and are not function. Lancet Lond. Engl. 2, 556–560 (1974). available due to data privacy laws. For HLA modeling, we used the 19. Iijima, K., Sako, M., Kamei, K. & Nozu, K. Rituximab in steroid- Protein Data Bank (PDB; https://www.rcsb.org/) and UNIPROT data- sensitive nephrotic syndrome: lessons from clinical trials. Pediatr. base (Uniprot ID: P01909). For colocalization analyses, we used the Nephrol. Berl. Ger. 33, 1449–1455 (2018). NEPTUNE cohort, the unfiltered eQTL results from DICE, and the 20. Gbadegesin, R. A. et al. HLA-DQA1 and PLCG2 are candidate risk loci BLUEPRINT consortium. for childhood-onset steroid-sensitive nephrotic syndrome. J. Am. Soc. Nephrol. JASN 26, 1701–1710 (2015). Code availability 21. Debiec, H. et al. Transethnic, genome-wide analysis reveals Publicly available software was used to perform the analyses. We will immune-related risk alleles and phenotypic correlates in pediatric be happy to share some or all our code with any requester. steroid-sensitive nephrotic syndrome. J. Am. Soc. Nephrol. JASN 29, 2000–2013 (2018). References 22. Jia, X. et al. Strong association of the HLA-DR/DQ locus with 1. Noone, D. G., Iijima, K. & Parekh, R. Idiopathic nephrotic syndrome in childhood steroid-sensitive nephrotic syndrome in the japanese children. Lancet 392, 61–74 (2018). population. J. Am. Soc. Nephrol. JASN 29, 2189–2199 (2018). 2. Gipson, D. S. et al. Gaining the PROMIS perspective from children 23. Jia, X. et al. Common risk variants in NPHS1 and TNFSF15 are asso- with nephrotic syndrome: a Midwest pediatric nephrology con- ciated with childhood steroid-sensitive nephrotic syndrome. Kidney sortium study. Health Qual. Life Outcomes 11, 30 (2013). Int. 98, 1308–1322 (2020). 3. Ruth, E. M., Landolt, M. A., Neuhaus, T. J. & Kemper, M. J. Health- 24. Dufek, S. et al. Genetic identification of two novel loci associated related quality of life and psychosocial adjustment in steroid- with steroid-sensitive nephrotic syndrome. J. Am. Soc. Nephrol. sensitive nephrotic syndrome. J. Pediatr. 145, 778–783 (2004). JASN 30, 1375–1384 (2019). Nature Communications | (2023)14:2481 9 Article https://doi.org/10.1038/s41467-023-37985-w 25. Julia, A. et al. A genome-wide association study identifies a novel 47. Li, X. et al. Genetic analyses identify GSDMB associated with asthma locus at 6q22.1 associated with ulcerative colitis. Hum. Mol. Genet. severity, exacerbations, and antiviral pathways. J. Allergy Clin. 23, 6927–6934 (2014). Immunol. 147, 894–909 (2021). 26. Schreiber, T. H. & Podack, E. R. Immunobiology of TNFSF15 and 48. Das, S., Miller, M. & Broide, D. H. Chromosome 17q21 Genes TNFRSF25. Immunol. Res. 57, 3–11 (2013). ORMDL3 and GSDMB in asthma and immune diseases. Adv. 27. Ovunc, B. et al. Mutation analysis of NPHS1 in a worldwide cohort of Immunol. 135 1–52 (2017). congenital nephrotic syndrome patients. Nephron Clin. Pract. 120, 49. Corces, M. R. et al. Lineage-specific and single-cell chromatin c139–c146 (2012). accessibility charts human hematopoiesis and leukemia evolution. 28. Mägi, R. et al. Trans-ethnic meta-regression of genome-wide asso- Nat. Genet. 48, 1193–1203 (2016). ciation studies accounting for ancestry increases power for dis- 50. Muto, Y. et al. Single cell transcriptional and chromatin accessibility covery and improves fine-mapping resolution. Hum. Mol. Genet. profiling redefine cellular heterogeneity in the adult human kidney. 26, 3639–3650 (2017). Nat. Commun. 12, 2190 (2021). 29. Ghoussaini, M. et al. Open Targets Genetics: systematic identifica- 51. Han, S. K. et al. Quality assessment and refinement of chromatin tion of trait-associated genes using large-scale genetics and func- accessibility data using a sequence-based predictive model. Proc, tional genomics. Nucleic Acids Res. 49, D1311–D1320 (2021). Natl. Acad. Sci. USA 119, e2212810119 (2022). 30. Ferland, R. J. et al. Abnormal cerebellar development and axonal 52. Luo, Y. et al. A high-resolution HLA reference panel capturing global decussation due to mutations in AHI1 in Joubert syndrome. Nat. population diversity enables multi-ethnic fine-mapping in HIV host Genet. 36, 1008–1013 (2004). response. https://doi.org/10.1101/2020.07.16.20155606 (2020). 31. Kukimoto-Niino, M. et al. Cryo-EM structure of the human ELMO1- 53. Frommer, L., Flesch, B. K., König, J. & Kahaly, G. J. Amino acid DOCK5-Rac1 complex. Sci. Adv. 7, eabg3147 (2021). polymorphisms in Hla class ii differentiate between thyroid and 32. Sharma, K. R. et al. ELMO1 protects renal structure and ultrafiltration polyglandular autoimmunity. J. Clin. Endocrinol. Metab. 105, in kidney development and under diabetic conditions. Sci. Rep. 6, dgz164 (2020). 37172 (2016). 54. Badenhoop, K. et al. Susceptibility and resistance alleles of human 33. Shimazaki, A. et al. Genetic variations in the gene encoding ELMO1 leukocyte antigen (HLA) DQA1 and HLA DQB1 are shared in endo- are associated with susceptibility to diabetic nephropathy. Diabetes crine autoimmune disease. J. Clin. Endocrinol. Metab. 80, 54, 1171–1178 (2005). 2112–2117 (1995). 34. Yu, C.-C. et al. Abatacept in B7-1–positive proteinuric kidney dis- 55. Rodrigues, C. H. M., Pires, D. E. V. & Ascher, D. B. DynaMut2: ease. N. Engl. J. Med. 369, 2416–2423 (2013). assessing changes in stability and flexibility upon single and mul- 35. Spada, R. et al. NKG2D ligand overexpression in lupus nephritis tiple point missense mutations. Protein Sci. Publ. Protein Soc. 30, correlates with increased NK cell activity and differentiation 60–69 (2021). in kidneys but not in the periphery. J. Leukoc. Biol. 97, 583–598 56. Ren, S. et al. Nephrotic syndrome associated with Kimura’s disease: (2015). a case report and literature review. BMC Nephrol. 19, 316 (2018). 36. Rayego-Mateos, S. et al. Role of epidermal growth factor receptor 57. Gallon, L., Leventhal, J., Skaro, A., Kanwar, Y. & Alvarado, A. Reso- (EGFR) and its ligands in kidney inflammation and damage. Media- lution of recurrent focal segmental glomerulosclerosis after tors Inflamm. 2018, 8739473 (2018). retransplantation. N. Engl. J. Med. 366, 1648–1649 (2012). 37. Ju, W. et al. Tissue transcriptome-driven identification of epidermal 58. Xu, X. et al. Molecular insights into genome-wide association stu- growth factor as a chronic kidney disease biomarker. Sci. Transl. dies of chronic kidney disease-defining traits. Nat. Commun. 9, Med. 7, 316ra193 (2015). 4800 (2018). 38. Rijvers, L. et al. The Role of Autoimmunity-Related Gene CLEC16A in 59. Neale Lab. Relationship of LDSR Results with Sample Size. UKB the B Cell Receptor-Mediated HLA Class II Pathway. J. Immunol. Heritability https://nealelab.github.io/UKBB_ldsc/viz_sampsize. Baltim. Md 1950 205, 945–956 (2020). html (2022). 39. Tam, R. C. Y. et al. Human CLEC16A regulates autophagy through 60. 1000 Genomes Project Consortium. et al. A global reference for modulating mTOR activity. Exp. Cell Res. 352, 304–312 (2017). human genetic variation. Nature 526, 68–74 (2015). 40. Pearson, G. et al. Clec16a, Nrdp1, and USP8 Form a Ubiquitin- 61. Wojcik, G. L. et al. Genetic analyses of diverse populations improves Dependent Tripartite Complex That Regulates β-Cell Mitophagy. discovery for complex traits. Nature 570, 514–518 (2019). Diabetes 67, 265–277 (2018). 62. Purcell, S. et al. PLINK: A Tool Set for Whole-Genome Association 41. Steimle, V., Siegrist, C. A., Mottet, A., Lisowska-Grospierre, B. & and Population-Based Linkage Analyses. Am. J. Hum. Genet. 81, Mach, B. Regulation of MHC class II expression by interferon- 559–575 (2007). gamma mediated by the transactivator gene CIITA. Science 265, 63. Taliun, D. et al. Sequencing of 53,831 diverse genomes from the 106–109 (1994). NHLBI TOPMed Program. Nature, 590, 290–299 (2021) 42. Edgar, A. J., Birks, E. J., Yacoub, M. H. & Polak, J. M. Cloning of 64. Das, S. et al. Next-generation genotype imputation service and dexamethasone-induced transcript: a novel glucocorticoid- methods. Nat. Genet. 48, 1284–1287 (2016). induced gene that is upregulated in emphysema. Am. J. Respir. Cell 65. Fuchsberger, C., Abecasis, G. R. & Hinds, D. A. minimac2: faster Mol. Biol. 25, 119–124 (2001). genotype imputation. Bioinformatics 31, 782–784 (2015). 43. Han, S. K. et al. Mapping genomic regulation of kidney disease 66. Hinrichs, A. S. et al. The UCSC Genome Browser Database: update and traits through high-resolution and interpretable eQTLs. Nat. 2006. Nucleic Acids Res. 34, D590–D598 (2006). Commun. 14, 2229 (2023). 67. Manichaikul, A. et al. Robust relationship inference in genome-wide 44. GTEx Consortium et al. Using an atlas of gene regulation across 44 association studies. Bioinformatics 26, 2867–2873 (2010). human tissues to inform complex disease- and trait-associated 68. Clarke, L. et al. The international Genome sample resource (IGSR): a variation. Nat. Genet. 50, 956–967 (2018). worldwide collection of genome variation incorporating the 1000 45. Schmiedel, B. J. et al. Impact of genetic polymorphisms on human Genomes Project data. Nucleic Acids Res. 45, D854–D859 (2017). immune cell gene expression. Cell 175, 1701–1715.e16 (2018). 69. Zhou, W. et al. Efficiently controlling for case-control imbalance 46. Fernández, J. M. et al. The BLUEPRINT data analysis portal. Cell Syst. and sample relatedness in large-scale genetic association studies. 3, 491–495.e5 (2016). Nat. Genet. 50, 1335–1341 (2018). Nature Communications | (2023)14:2481 10 Article https://doi.org/10.1038/s41467-023-37985-w 70. Abraham, G., Qiu, Y. & Inouye, M. FlashPCA2: principal component Acknowledgements analysis of Biobank-scale genotype datasets. Bioinformatics Oxf. K.I., K.N., and K.T. were supported by the Japan Agency for Medical Engl. 33, 2776–2778 (2017). Research and Development (AMED) under grant number 71. Delaneau, O., Marchini, J. & Zagury, J.-F. A linear complexity phas- JP17km0405108h0005. K.T. was supported by the Japan Agency ing method for thousands of genomes. Nat. Methods 9, for Medical Research and Development (AMED) under grants 179–181 (2011). JP17km0405205h0002 and 18km0405205h0003. K.I., T.H., C.N., and 72. Howie, B. N., Donnelly, P. & Marchini, J. A flexible and accurate K.N. were supported by the Japan Society for the Promotion of Science genotype imputation method for the next generation of genome- (JSPS) under Grant-in-Aid for Scientific Research fostering Joint Inter- wide association studies. PLoS Genet. 5, e1000529 (2009). national Research (B) 18KK0244. K.I., X.J., T.H., C.N., and K.N. were 73. Willer, C. J., Li, Y. & Abecasis, G. R. METAL: fast and efficient meta- supported by the Japan Society for the Promotion of Science (JSPS) analysis of genomewide association scans. Bioinformatics 26, under Grant-in-Aid for Scientific Research fostering Joint International 2190–2191 (2010). Research (B) 21KK0147. This work is supported by the Department of 74. Pruim, R. J. et al. LocusZoom: regional visualization of genome-wide Defense (PR190746, PR212415) to S.S-C., by the National Center for association scan results. Bioinformatics 26, 2336–2337 (2010). Advancing Translational Sciences, National Institutes of Health (Grant 75. Yang, J., Lee, S. H., Goddard, M. E. & Visscher, P. M. GCTA: a tool for Number UL1TR001873) to S.S-C., and by the National Institute of Health genome-wide complex trait analysis. Am. J. Hum. Genet. 88, Grant RC2DK122397, M.Sam, S.S-C., M.R.P., and F.H. A.M. received 76–82 (2011). support from the American Society of Nephrology KidneyCure Ben J. 76. Yang, J. et al. Conditional and joint multiple-SNP analysis of GWAS Lipps Research Fellowship. Y.G. received support from the NEPTUNE summary statistics identifies additional variants influencing com- Career Development Award. P.R. and H.D. were funded by European plex traits. Nat. Genet. 44, 369–375 (2012). Research Council grant ERC-2012- ADG_20120314 (grant agreement 77. Bulik-Sullivan, B. K. et al. LD Score regression distinguishes con- 322947) and Agence Nationale pour la Recherche “Genetransnephrose” founding from polygenicity in genome-wide association studies. grant ANR-16-CE17-004-01. M.Sam. was supported by NIH grants Nat. Genet. 47, 291–295 (2015). R01DK119380, 2U54DK083912, and a gift from The Pura Vida Kidney 78. Wen, X., Pique-Regi, R. & Luca, F. Integrating molecular QTL data Foundation. The Nephrotic Syndrome Study Network (NEPTUNE) is part into genome-wide genetic association analysis: probabilistic of the Rare Diseases Clinical Research Network (RDCRN), which is fun- assessment of enrichment and colocalization. PLOS Genet. 13, ded by the National Institutes of Health (NIH) and led by the National e1006646 (2017). Center for Advancing Translational Sciences (NCATS) through its Divi- 79. Gadegbeku, C. A. et al. Design of the Nephrotic Syndrome Study sion of Rare Diseases Research Innovation (DRDRI). NEPTUNE is funded Network (NEPTUNE) to evaluate primary glomerular nephropathy under grant number U54DK083912 as a collaboration between NCATS by a multidisciplinary approach. Kidney Int. 83, 749–756 (2013). and the National Institute of Diabetes and Digestive and Kidney Diseases 80. Wen, X., Lee, Y., Luca, F. & Pique-Regi, R. Efficient integrative multi- (NIDDK). Additional funding and/or programmatic support is provided SNP association analysis via deterministic approximation of pos- by the University of Michigan, NephCure Kidney International, and the teriors. Am. J. Hum. Genet. 98, 1114–1129 (2016). Halpin Foundation. RDCRN consortia are supported by the RDCRN Data 81. Shi, H. et al. Localizing components of shared transethnic genetic Management and Coordinating Center (DMCC), funded by NCATS and architecture of complex traits from GWAS summary data. Am. J. the National Institute of Neurological Disorders and Stroke (NINDS) Hum. Genet. 106, 805–817 (2020). under U2CTR002818. The authors wish to thank Seong Kyu Han, Ph.D. 82. Maller, J. B. et al. Bayesian refinement of association signals for 14 (Boston Children’s Hospital and Harvard Medical School) for his assis- loci in 3 common diseases. Nat. Genet. 44, 1294–1301 (2012). tance in creating figures. 83. Ting, Y. T. et al. A molecular basis for the T cell response in HLA- DQ2.2 mediated celiac disease. Proc. Natl Acad. Sci. USA 117, 3063–3073 (2020). Author contributions 84. Mooers, B. H. M. Shortcuts for faster image creation in PyMOL. M.Sam., K.I., P.R., S.S-C., K.T. S.R., Y.L., K.K., and D.L. devised the main Protein Sci. Publ. Protein Soc. 29, 268–276 (2020). conceptual idea of this project. A.Bar, M.Mc., Y.G., M.Sam., K.I., P.R., S.S- 85. Halgren, T. A. MMFF VII. Characterization of MMFF94, MMFF94s, C., K.T S wrote the paper. A.Bar, M.Mc, X.J., S.J., C.N., Y.K., K.S., K.T., Y.G., and other widely available force fields for conformational energies D.F.A., G.J., F.Z., and T.Y.L. conducted statistical analysis and visualiza- and for intermolecular-interaction energies and geometries. J. tion. K.N., C.N., T.H., H.II.C, P.V., and K.I. collected and managed patient Comput. Chem. 20, 730–748 (1999). specimens and clinical information from each country with the coop- 86. Johansson, M. U., Zoete, V., Michielin, O. & Guex, N. Defining and eration of pediatric nephrologists in Japan, Korea, and Thailand listed in searching for structural motifs using DeepView/Swiss-PdbViewer. the Supplemental Materials. P.R. and H.D. coordinated European and BMC Bioinformatics 13, 173 (2012). Indian cohorts, reviewed the phenotypes of the patients and their out- 87. Pol-Fachin, L., Fernandes, C. L. & Verli, H. GROMOS96 43a1 per- comes, and collected, and genotyped the samples at Integragen (Evry, formance on the characterization of glycoprotein conformational France). The following authors recruited and phenotyped patients: M.V., ensembles through molecular dynamics simulations. Carbohydr. M.C., G.M., W.M., A.Bag, A.S., M.Sc, N.T., A.Z., M.D-D., C.D., J.H., G.A., Res. 344, 491–500 (2009). A.J., L.P., A.M., G.C., M.B., R.W., M.Ma., D.T., M.G., L.G., R.M., I.P., E.F., J.R., 88. Jubb, H. C. et al. Arpeggio: a web server for calculating and S.M., P.L.W., F.S., M.Sar., V.T., D.S., J.A.E.VW., D.M., M.R.P., A.G., F.L., visualising interatomic interactions in protein structures. J. Mol. A.C.SeS., R.J.F.L., E.E.K, F.H., S.S., G.M.G., and S.S-C. All authors Biol. 429, 365–371 (2017). approved the submission of the final version of the article for 89. Ruan, Y. et al. Improving polygenic prediction in ancestrally diverse publication. populations. Nat. Genet. 54, 573–580 (2022). 90. Ge, T., Chen, C.-Y., Ni, Y., Feng, Y.-C. A. & Smoller, J. W. Polygenic prediction via Bayesian regression and continuous shrinkage priors. Competing interests Nat. Commun. 10, 1776 (2019). The authors declare no competing interests. Nature Communications | (2023)14:2481 11 Article https://doi.org/10.1038/s41467-023-37985-w Additional information Open Access This article is licensed under a Creative Commons Supplementary information The online version contains Attribution 4.0 International License, which permits use, sharing, supplementary material available at adaptation, distribution and reproduction in any medium or format, as https://doi.org/10.1038/s41467-023-37985-w. long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if Correspondence and requests for materials should be addressed to changes were made. The images or other third party material in this Matthew G. Sampson. article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not Peer review information Nature Communications thanks the anon- included in the article’s Creative Commons license and your intended ymous reviewers for their contribution to the peer review of this work. use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright Reprints and permissions information is available at holder. To view a copy of this license, visit http://creativecommons.org/ http://www.nature.com/reprints licenses/by/4.0/. Publisher’s note Springer Nature remains neutral with regard to jur- © The Author(s) 2023 isdictional claims in published maps and institutional affiliations. Alexandra Barry1,2,59, Michelle T. McNulty 1,2,59, Xiaoyuan Jia3,4,59, Yask Gupta5,59, Hanna Debiec6,59, Yang Luo 7,8,9,10, China Nagano 1,2,11, Tomoko Horinouchi 11, Seulgi Jung 12, Manuela Colucci13, Dina F. Ahram5, Adele Mitrotti5,14, Aditi Sinha15, Nynke Teeninga16, Gina Jin5, Shirlee Shril17,18, Gianluca Caridi 19, Monica Bodria20, Tze Y. Lim5, Rik Westland 21, Francesca Zanoni5,22, Maddalena Marasa 5, Daniel Turudic23, Mario Giordano24, Loreto Gesualdo14, Riccardo Magistroni 25,26, Isabella Pisani27, Enrico Fiaccadori27, Jana Reiterova28, Silvio Maringhini29, William Morello 30, Giovanni Montini 30,31, Patricia L. Weng32, Francesco Scolari33, Marijan Saraga34, Velibor Tasic 35, Domenica Santoro36, Joanna A. E. van Wijk21, Danko Milošević23,37, Yosuke Kawai 3,4, Krzysztof Kiryluk5, Martin R. Pollak38,39, Ali Gharavi 5, Fangmin Lin 39, Ana Cristina Simœs e Silva 40, Ruth J. F. Loos 41, Eimear E. Kenny42,43,44, Michiel F. Schreuder 16, Aleksandra Zurowska45, Claire Dossier46, Gema Ariceta 47, Magdalena Drozynska-Duklas45, Julien Hogan46, Augustina Jankauskiene48, Friedhelm Hildebrandt 1,18, Larisa Prikhodina 49, Kyuyoung Song 12, Arvind Bagga15, Hae Cheong II 50, Gian Marco Ghiggeri 20, Prayong Vachvanichsanong51, Kandai Nozu 11, Dongwon Lee 1,2,18, Marina Vivarelli 52, Soumya Raychaudhuri 8,9,10,53,54, Katsushi Tokunaga 3,4,60, Simone Sanna-Cherchi 5,60, Pierre Ronco 6,55,60, Kazumoto Iijima56,57,60 & Matthew G. Sampson 1,2,18,58,60 1 Division of Nephrology, Boston Children’s Hospital, Boston, MA, USA. 2Kidney Disease Initiative & Medical and Population Genetics Program, Broad Institute of MIT and Harvard, Cambridge, MA, USA. 3Genome Medical Science Project (Toyama), National Center for Global Health and Medicine (NCGM), Tokyo, Japan. 4 Department of Human Genetics, Graduate School of Medicine, The University of Tokyo, Tokyo, Japan. 5Division of Nephrology, Department of Medicine, Columbia University College of Physicians and Surgeons, New York, NY, USA. 6Sorbonne Université, UPMC Paris 06, Institut National de la Santé et de la Recherde Médicale, Unité Mixte de Rechereche, S 1155 Paris, France. 7Kennedy Institute of Rheumatology, University of Oxford, Roosevelt Drive, Headington, Oxford OX3 7FY, United Kingdom. 8Center for Data Sciences, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA. 9Divisions of Genetics and Rheumatology, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA. 10Program in Medical and Population Genetics, Broad Institute of MIT and Harvard, Cambridge, MA, USA. 11Department of Pediatrics, Kobe University Graduate School of Medicine, Kobe, Japan. 12Department of Biochemistry and Molecular Biology, University of Ulsan College of Medicine, Songpa-gu, Seoul, Korea. 13Renal Diseases Research Unit, Genetics and Rare Diseases Research Division, Istituto di Ricovero e Cura a Carattere Scientifico Ospedale Pediatrico Bambino Gesù, Rome, Italy. 14Nephrology, Dialysis and Transplantation Unit, Department of Emergency and Organ Transplantation, University of Bari Aldo Moro, Bari, Italy. 15 Department of Pediatrics, AIIMS, New Delhi, India. 16Department of Pediatric Nephrology, Amalia Children’s Hospital, Radboud University Medical Center, Nijmegen, The Netherlands. 17Department of Medicine, Boston Children’s Hospital, Boston, MA, USA. 18Department of Pediatrics, Harvard Medical School, Boston, MA, USA. 19Laboratory on Molecular Nephrology, IRCCS Instituto Giannina Gaslini, Genoa, Italy. 20Department of Nephrology and Renal Trans- plantation, IRCCS Instituto Giannina Gaslini, Genoa, Italy. 21Department of Pediatric Nephrology, VU University Medical Center, Amsterdam, The Netherlands. 22 Division of Transplantation, Department of Surgery, University of Pennsylvania, Philadelphia, PA, USA. 23Department of Pediatric Nephrology, Dialysis and Transplantation, Clinical Hospital Hospital Center Zagreb, University of Zagreb Medical School, Zagreb, Croatia. 24Division of Nephrology and Pediatric Dialysis, Bari Polyclinic Giovanni XXIII Children’s Hospital, Bari, Italy. 25Department of Nephrology, Dialysis and Transplant Unit, University Hospital of Modena, Modena, Italy. 26Surgical, Medical and Dental Department of Morphological Sciences, Section of Nephrology, University of Modena and Reggio Emilia, Modena, Italy. 27Unità Operativa Nefrologia, Azienda Ospedaliero-Universitaria di Parma, Dipartimento di Medicina e Chirurgia, Università di Parma, Parma, Italy. 28Department of Nephrology, Medicine and General University Hospital, Charles University, Prague, Czech Republic. 29Department of Pediatrics, ISMETT, Palermo, Italy. 30Pediatric Nephrology, Dialysis and Transplant Unit, Fondazione IRCCS Ca’ Granda-Ospedale Maggiore Policlinico, Milano, Italy. 31 Department of Clinical Sciences and Community Health, University of Milan, Milan, Italy. 32Department of Pediatric Nephrology, UCLA Medical Center and UCLA Medical Center-Santa Monica, Los Angeles, CA, USA. 33Department of Medical and Surgical Specialties, Radiological Sciences, and Public Health, Division of Nephrology and Dialysis, University of Brescia and ASST Spedali Civili of Brescia, Brescia, Italy. 34Department of Pediatrics, University of Split, Split, Croatia. 35Department of Pediatric Nephrology, University Children’s Hospital, Skopje, Macedonia. 36Division of Nephrology and Dialysis Unit, University of Messina, Sicily, Italy. 37Croatian Academy of Medical Sciences, Praska 2/III p.p. 27, 10000 Zagreb, Croatia. 38Division of Nephrology, Beth Israel Deaconess Nature Communications | (2023)14:2481 12 Article https://doi.org/10.1038/s41467-023-37985-w Medical Center, Boston, MA, USA. 39Department of Pediatric, Division of Pediatric Nephrology, Columbia University Irving Medical Center New York- Presbyterian Morgan Stanley Children’s Hospital in New York, New York, NY, USA. 40Department of Pediatrics, Interdisciplinary Laboratory of Medical Investigation, Faculty of Medicine, Federal University of Minas Gerais, Belo Horizonte, Brazil. 41The Charles Bronfman Institute for Personalized Medicine, Icahn School of Medicine at Mount Sinai, New York, NY, USA. 42Institute for Genomic Health, Icahn School of Medicine at Mount Sinai, New York, NY, USA. 43Division of Genomic Medicine, Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, NY, USA. 44Division of General Internal Medicine, Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, NY, USA. 45Department of Pediatrics, Nephrology and Hypertension, Medical University Gdansk, Gdansk, Poland. 46AP-HP, Pediatric Nephrology Department, Hôpital Robert-Debré, Paris, France. 47Pediatric Nephrology, Hospital Uni- versitari Vall d’Hebron, Universitat Autónoma de Barcelona, Barcelona, Spain. 48Institute of Clinical Medicine, Faculty of Medicine, Vilnius University, Vilnius, Lithuania. 49Research and Clinical Institute for Pediatrics, Pirogov Russian National Research Medical University, Taldomskava St, 2, Moscow, Russia. 50 Department of Pediatrics, Hallym University Sacred Heart Hospital, 22, Gwanpyeong-ro 170 beon-gil, Dongan-gu, Anyang-si, Gyeonggi-do 14068, Korea. 51 Department of Pediatrics, Faculty of Medicine, Prince of Songkla University, Hat-Yai, Songkhla 90110, Thailand. 52Division of Nephrology, and Dialysis, Department of Pediatric Subspecialities, Istituto di Ricovero e Cura a Carattere Scientifico Ospedale Pediatrico Bambino Gesù, Rome, Italy. 53Department of Biomedical Informatics, Harvard Medical School, Boston, MA, USA. 54Centre for Genetics and Genomics Versus Arthritis, University of Manchester, Manchester, UK. 55Department of Nephrology, Centre Hospitalier du Mans, Le Mans, France. 56Hyogo Prefectural Kobe Children’s Hospital, Kobe, Japan. 57 Department of Advanced Pediatric Medicine, Kobe University Graduate School of Medicine, Kobe, Japan. 58Division of Renal Medicine, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA. 59These authors contributed equally: Alexandra Barry, Michelle T. McNulty, Xiaoyuan Jia, Yask Gupta, Hanna Debiec. 60These authors jointly supervised this work: Katsushi Tokunaga, Simone Sanna-Cherchi, Pierre Ronco, Kazumoto Iijima, Matthew G. Sampson. e-mail: matthew.sampson@childrens.harvard.edu Nature Communications | (2023)14:2481 13