|Home | About | Journals | Submit | Contact Us | Français|
African Americans (AAs) have increased susceptibility to non-diabetic nephropathy relative to European Americans.
Follow-up of a pooled genome-wide association study (GWAS) in AA dialysis patients with nondiabetic nephropathy; novel gene-gene interaction analyses.
Wake Forest sample: 962 AA nondiabetic nephropathy cases; 931 non-nephropathy controls. Replication sample: 668 Family Investigation of Nephropathy and Diabetes (FIND) AA nondiabetic nephropathy cases; 804 non-nephropathy controls.
Individual genotyping of top 1420 pooled GWAS-associated single nucleotide polymorphisms (SNPs) and 54 SNPs in six nephropathy susceptibility genes.
APOL1 genetic association and additional candidate susceptibility loci interacting with, or independently from, APOL1.
The strongest GWAS associations included two non-coding APOL1 SNPs, rs2239785 (odds ratio [OR], 0.33; dominant; p = 5.9 × 10−24) and rs136148 (OR, 0.54; additive; p = 1.1 × 10−7) with replication in FIND (p = 5.0 × 10−21 and 1.9 × 10−05, respectively). Rs2239785 remained significantly associated after controlling for the APOL1 G1 and G2 coding variants. Additional top hits included a CFH SNP(OR from meta-analysis in above 3367 AA cases and controls, 0.81; additive; p = 6.8 × 10−4). The 1420 SNPs were tested for interaction with APOL1 G1 and G2 variants. Several interactive SNPs were detected, the most significant was rs16854341 in the podocin gene (NPHS2) (p = 0.0001).
Non-pooled GWAS have not been performed in AA nondiabetic nephropathy.
This follow-up of a pooled GWAS provides additional and independent evidence that APOL1 variants contribute to nondiabetic nephropathy in AAs and identified additional associated and interactive non-diabetic nephropathy susceptibility genes.
Relative to European Americans (EAs), African Americans (AAs) have increased susceptibility to nondiabetic forms of end-stage renal disease (ESRD), an effect that persists after controlling for hypertension and socioeconomic factors1. Although family studies in AAs confirm a strong genetic component to nondiabetic ESRD2;3, few studies have focused on identifying susceptibility genes in AAs.
Recent studies implicated the chromosome 22q region as significantly associated with nondiabetic chronic kidney disease (CKD) in AAs 4–6. Genovese et al. recently demonstrated highly significant association of mutations in the apolipoprotein L1 (APOL1) gene (the “G1” allele, consisting of reference single-nucleotide polymorphism [SNP] identification number 73885319 [rs73885319] and rs60910145 in perfect linkage disequilibrium [r2=1.0 and tagged by either SNP] and the “G2” allele, corresponding to rs71785313) in AA cases with biopsy-proven focal segmental glomerulosclerosis (FSGS) and dialysis patients with nondiabetic nephropathy 7;8. Evidence of association remained significant after controlling for other associated SNPs in this region, including in the non-muscle myosin heavy chain 9 gene (MYH9).
Investigators at the Wake Forest School of Medicine performed a genome-wide association study (GWAS) for nondiabetic nephropathy in AAs on dialysis using pooled DNA with follow-up genotyping of 62 SNPs in individual cases and controls 9. The present study reports the admixture-adjusted results of additional individual genotyping of the top 1420 SNPs identified from this GWAS in AA nondiabetic nephropathy cases and non-nephropathy controls, with replication in a similar cohort of AAs with and without nondiabetic nephropathy. Study objectives were to confirm association identified in the APOL1 gene region on chromosome 22 and identify additional candidate nondiabetic nephropathy susceptibility loci in AAs that may interact with, or act independently from, APOL1.
Wake Forest nondiabetic nephropathy cases were recruited from dialysis centers in North Carolina, South Carolina, Georgia, Virginia and Tennessee. Participants were self-described AA dialysis patients lacking diabetes mellitus at initiation of renal replacement therapy. Patients were classified as having nondiabetic nephropathy if they had hypertension or a primary or secondary chronic glomerular disease listed as their cause of nephropathy. Those classified as cases confirmed the onset of high blood pressure prior to dialysis initiation, in the absence of other kidney disease risk factors. Hypertension-attributed nephropathy was diagnosed in the presence of proteinuria ≤ 1.5 gm/day, urinalysis ≤ 100 mg/dl protein, or spot urine protein-creatinine ratio ≤ 1.5 gm/gm (when measures were available); typically with evidence of other hypertensive target organ damage. Chronic glomerulonephritis was diagnosed in those with kidney biopsy evidence or proteinuria ≥ 1.5 gm/day. Patients with polycystic kidney disease, Alport’s syndrome, IgA nephropathy, urologic disease or surgical nephrectomy were excluded. Controls were self-described healthy AAs over the age of 18 years recruited from community sources and medical clinics in North Carolina who denied a history of kidney disease. Subjects provided blood for DNA extraction after providing written informed consent. Study procedures were in compliance with the Wake Forest School of Medicine Institutional Review Board and Declaration of Helsinki.
The replication DNA samples for the top 3 pooled GWAS hits came from the Family Investigation of Nephropathy and Diabetes (FIND) AA nondiabetic nephropathy Mapping by Admixture Linkage Disequilibrium (MALD) study (545 cases and 804 controls) and the CHOICE (Choices for Healthy Outcomes In Caring for ESRD) study, which is a national prospective cohort study of 1041 participants on dialysis (including 123 with nondiabetic nephropathy). FIND and CHOICE participant details have previously been published 5;10, Of FIND and CHOICE cases (hereafter referred to simply as FIND), 347 had hypertension-attributed nephropathy, 87 FSGS, 69 HIV-associated nephropathy, 126 other glomerular disorders and 39 other causes. AA FIND controls had estimated glomerular filtration rates > 60 ml/min/1.73 m2 and spot urinary albumin-creatinine ratios < 30 mg/gm.
To identify genes which contribute to non-diabetic nephropathy, we performed a GWAS using pooled DNA 9. DNA from 500 AA dialysis patients with nondiabetic nephropathy (cases) and 500 non-nephropathy controls were divided into 10 pools each and genotyped on the Illumina HumanHap 500 genotyping chip. The 65 most highly associated SNPs were genotyped on the same 1000 samples as individual DNAs. Thirty-six nephropathy cases and 22 controls were eliminated from this analysis due to failed genotyping or missing phenotype data. These 65 SNPs were replicated on 336 additional nondiabetic nephropathy cases and 363 non-nephropathy controls. Sixteen SNPs were associated with p-values < 7.7 × 10−4, 12 of which were located on chromosome 22, in or around the MYH9/APOL1 gene region. Several SNPs in MYH9 were strongly associated with nondiabetic nephropathy (P values ranging from 0.027 to 2.601 × 10−14)9. SNPs from the pooled DNA analysis were selected as autosomal SNPs with association p-values < 0.001. The initially identified SNPs in the MYH9 gene, which were highly associated with nondiabetic nephropathy and in strong linkage disequilibrium with APOL1, were excluded since they were evaluated in a prior report6, as were those with minor allele frequencies < 0.05 in the HapMap sampling Yoruba in Ibadan (YRI), Nigeria (www.hapmap.org). SNPs associated with nondiabetic nephropathy after analysis of the Wake Forest pooled GWAS and SNPs associated with nephropathy in a subset of AASK (African American Study of Kidney Disease and Hypertension) participants undergoing a preliminary GWAS at the Mount Sinai School of Medicine (p < 0.01) were included (courtesy of Drs. Iyengar, Lipkowitz and Bottinger). An additional 54 tag SNPs were evaluated in six genes known to be associated with FSGS or nephropathy: α actinin-4 (ACTN4) 11, nephrin (NPHS1)12, podocin (NPHS2)13, calcium-permeable canonical transient receptor potential 6 (TRPC6)14, uromodulin (UMOD)15, the amino acid transporter SLC7A916, and neurocalcin delta (NCALD)17. In total, 1536 SNPs were selected for genotyping (including 44 blind duplicates) on 988 Wake Forest nondiabetic nephropathy cases and 1036 controls (including the 500 Wake Forest cases and 500 controls from the original pooled GWAS) on the Illumina GoldenGate custom genotyping chip by the Center for Inherited Disease Research at Johns Hopkins University.
To adjust for African ancestry in Wake Forest samples, 70 ancestry-informative markers were also genotyped on all nondiabetic nephropathy cases and controls and the percentage of African ancestry was calculated, as described 18. HapMap 3 (http://hapmap.ncbi.nlm.nih.gov/) genotypes from the YRI, CEU (Utah residents with Northern and Western European ancestry from the CEPH [Centre de’Etude du Polymorphism Humain] collection), and HAN (Han Chinese) were used to anchor these calculations.
The three most associated SNPs (two in APOL1 and one in the complement factor H gene [CFH]) and 1 additional SNP in MYH9 (rs4821480) were genotyped in a replication sample from the FIND African American MALD study. The top SNPs were chosen because they were located in the APOL1 gene, as published previously; other top ranking SNPs were also located on chromosome 22 and are in linkage disequilibrium with the more strongly associated SNPs in APOL1. The SNP located in the CFH gene was the highest-ranking SNP not located on chromosome 22. In addition, CFH has been shown to have functional significance in kidney disease. The three candidate SNPs were genotyped using the ABI Taqman platform. Ancestry informative markers (N=1354) were genotyped in FIND using the Illumina GoldenGate custom genotyping chip 5.
Comparison of demographic measures between Wake Forest cases and controls were calculated using a Wilcoxon rank sum test (SigmaStat 3.5, Systat Software, San Jose, CA). Tests for Hardy-Weinberg Equilibrium and genotypic association as well as associated summary statistics were computed using the analysis program SNPGWA 19;20. Specifically, tests of association were computed adjusting for age, gender and percent African ancestry using a logistic regression model. Tests of association were computed for dominant, additive and recessive genetic models as well as a lack-of-fit to an additive genetic model. The primary inference is the additive genetic model unless the lack-of-fit test was significant (p<0.05). In this case, the minimum p-value of the dominant, additive, and recessive genetic models is reported. Local ancestry for the MYH9 region of chromosome 22 was calculated using 14 SNPs from the MYH9 gene which were identified as part of a genome-wide admixture scan for non-diabetic nephropathy. Local ancestry was calculated at each marker separately, and these 14 estimates were used to adjust the association analysis as in Kao et al 5. To account for the number of hypotheses tested (i.e., multiple comparisons), a false discovery rate–adjusted p-value was computed for 904 comparisons. The latter number was obtained from a principal component analysis on the genotype data where 904 principal components explained >99% of the genetic variation in the SNPs 21;22; thus, due to linkage disequilibrium there are approximately 904 independent dimensions in the SNP data and therefore 904 independent hypotheses.
The three SNPs tested for replication in the FIND were analyzed for association with nondiabetic nephropathy in unrelated AA cases and controls as described above while adjusting for age, gender, global and local ancestries, which were assessed as percent European ancestry. Global ancestry was estimated with 1354 markers and local with 23 markers using the software ANCESTRYMAP23.
To account for the recently described highly significant association of the APOL1 gene (G1 [rs73885319 and rs60910145] and G2 [rs71785313] risk variants) with nondiabetic kidney disease in AAs 7, we completed two sets of analyses under a logistic regression model framework. Because of the linkage disequilibrium pattern where G1 and G2 risk variants are very rarely observed together, we constructed a binary variable representing the compound G1/G2 risk across these three markers, modeling APOL1 risk as the response for all individuals with recessive haplotypes at either G1 or G2 or heterozygosity at both G1 and G2. To test for an interaction between the APOL1 risk loci and each SNP we computed a logistic regression model with the G1/G2 compound risk variable and the individual SNP as covariates and tested the interaction as modeled using the standard centered cross-product of G1/G2 and the SNP. In addition, we computed the logistic regression model without the interaction term to test for the association of each SNP after accounting for the effects of G1/G2. These two sets of analyses continue to adjust for age, gender and percentage of African ancestry.
In Wake Forest samples (962 nondiabetic nephropathy cases; 931 non-nephropathy controls), adjusting for the number of comparisons and assuming an additive genetic model with a risk allele frequency of 0.20 and a type 1 error rate of 0.0001, the power to detect ORs of 1.35 and 1.45 was ~0.50 and ~0.80, respectively. Parallel power calculations for validating the three most associated SNPs in the FIND samples (668 nondiabetic nephropathy cases; 804 non-nephropathy controls), but for a type 1 error rate of 0.05, revealed the power to detect ORs of 1.19 and 1.29 was ~0.50 and ~0.80, respectively.
Among Wake Forest participants, nondiabetic nephropathy cases were older, had higher BMI, fewer females, and greater African ancestry, compared to non-nephropathy controls (Table 1). In FIND participants, nondiabetic nephropathy cases were slightly older, had fewer females, and lower BMI compared to controls.
Of the 1536 SNPs that were selected for genotyping, 1420 were successful and used in these analyses. The mean SNP call rate was 0.999 (standard deviation 0.001). Of the 1980 DNA samples sent to the Center for Inherited Disease Research at Johns Hopkins University for genotyping, 1846 were retained for analysis after removal of failed or poorly genotyped samples, unexpected duplicates and unexpected relatedness. There was no evidence that individuals removed from the analysis based on quality control analysis differed from those analyzed for BMI, age at hypertension, age at dialysis, age at enrollment, admixture or gender (p>0.20). Phenotype data were missing in the other 87 excluded participants. Genotyping rates by individual ranged from 95.5% to 100%; less than 1% had genotyping rates below 98%.
Association analysis in Wake Forest participants genotyped on the GoldenGate chip resulted in 61 SNPs associated at p-values < 0.01 (Table 2). The two most associated SNPs, rs2239785 and rs136148, were in the APOL1 gene, (p = 5.91 × 10−24 [dominant] and p = 1.13 × 10−7 [additive]), consistent with recently published results 7. Two additional SNPs were associated with p-values < 3.00 × 10−5 (remaining significantly associated after Bonferroni correction), rs1573708 in an intergenic region and rs4820237 in the FOXRED2 gene, both located on chromosome 22. Additional SNPs of interest were rs379489 in the complement factor H gene (CFH) (p = 2.05 × 10−4, additive; OR, 0.73) and rs16854341 in the podocin gene (NPHS2; p = 0.004, dominant; OR, 0.72), as both genes have previously been associated with kidney disease 24;25. In addition, the APOL1 G1 and G2 variants were strongly associated with nephropathy in these Wake Forest cases (p = 3.27 × 10−32, recessive; OR, 5.21; 95% CI, 3.96–6.86).
To address the issue of multiple comparisons, we estimated the number of independent tests performed. A principal component analysis computed on the 1420 SNPs identified 904 principal components, suggesting 904 independent hypotheses tested. False discovery rate–adjusted p-values suggest that besides the chromosome 22 loci, the CFH rs379489 locus was significant (false discovery rate–adjusted p-value=0.03).
The top two APOL1 SNPs from the association analysis, rs2239785 and rs136148, were tested for replication in 1474 FIND participants (668 nondiabetic nephropathy cases, 804 controls), FIND participants were not included in the initial reports of APOL1 association 7;8. Both SNPs replicated, with association p-values of 5.03 × 10−21 and 1.92 × 10−5, respectively (Table 3). Adjustment for local ancestry on chromosome 22 did not significantly alter the significance of either APOL1 SNP; rs2239785 and rs136148 remained strongly associated after adjustment for local admixture. Importantly, the MYH9 E1 haplotype SNP rs4821480 (p = 1.12 × 10−10) remained strongly associated with nondiabetic nephropathy after adjusting for local ancestry and these APOL1 SNPs.
We attempted to replicate the CFH SNP rs379489 association with nondiabetic nephropathy in the FIND participants. Although the association between this SNP was similar between the FIND and Wake Forest participants, the association was non-significant in FIND alone (p = 0.345; odds ratio, 0.92; additive). When FIND and Wake Forest samples were combined in a meta-analysis, rs379489 remained significantly associated (p = 6.75 × 10−4; odds ratio, 0.81; additive); although association was driven primarily by Wake Forest samples. The heterogeneity p-value in the meta-analysis was 0.0769 (additive).
As APOL1 G1 and G2 variants are known to be strongly associated with nondiabetic nephropathy in AAs, we tested the 1420 SNPs that were genotyped for interaction with the G1 and G2 nephropathy risk variants (Table 4). The most significantly associated interactive SNP was rs16854341 on chromosome 1 in the podocin (NPHS2) gene (p = 0.0001). NPHS2 is independently associated with susceptibility to glomerulosclerosis 13;25–27. Six other SNPs were associated with p-values < 0.001; however, none were located in known kidney disease candidate genes.
In order to increase the statistical power for association, an exploratory analysis was computed using logistic regression models that tested for association with each of the 1420 SNPs adjusting for the APOL1 G1/G2 risk loci (Table 5). Most loci provided comparable evidence with and without adjustment for the G1/G2 risk loci. However, five loci showed meaningful improvement in the association: rs1500474 on 2q37 (unadjusted OR, 0.94 [p=0.39]; adjusted OR, 0.61 [p=0.0018]); rs11191727 on 10q24 within the gene neuralized-like protein 1 (NEURL1; unadjusted OR, 0.73 [p=0.012]; adjusted OR, 0.59 [p=0.00033]); rs1355652 on 11q14 (unadjusted OR, 1.36 [p=0.0028]; adjusted OR, 1.58 [p=0.00014]); rs9318258 on 13q22 (unadjusted OR, 0.82 [p=0.0070]; adjusted OR, 0.75 [ p=0.00070]) and rs10483956 on 14q31 (unadjusted OR, 0.76 [p=0.0056]; adjusted OR, 0.68 [p=0.00074]). The novel APOL1 SNP, rs2239785, remained modestly associated (p=0.00363). This SNP also remained associated in the FIND population when adjusted for G1/G2 risk status (p=0.0000518; OR, 0.52) (Table 3).
The present report confirmed and replicated the top SNP associations from a pooled GWAS for non-diabetic nephropathy in AAs 9 in a large sample of nondiabetic nephropathy cases and non-nephropathy controls. It also reports the first APOL1 G1/G2 gene-gene interaction analysis including all of the top GWAS SNPs, plus 54 additional tag SNPs in six glomerulosclerosis-associated genes. Two novel SNPs in APOL1 were robustly associated with nondiabetic nephropathy in AAs, a result consistent with recently published identification of the G1 and G2 risk loci in these 7 and other populations8. We successfully replicated association of these new APOL1 SNPs in another large sample of AA nondiabetic nephropathy cases and healthy non-CKD controls from FIND. When these SNPs were tested for association after adjustment for rs136148 in APOL1 and an MYH9 SNP in the E1 haplotype rs48214804;5, rs2239785 remained strongly associated (p = 9.14 × 10−13). Rs2239785 remained nominally associated with nondiabetic nephropathy after adjustment for the known coding G1 and G2 alleles in APOL1 (p = 3.63 × 10−3 in Wake Forest samples; p = 5.18 × 10−5 in FIND samples). Adjustment for local ancestry on chromosome 22 did not significantly change association at this SNP (p = 1.12 × 10−10). In addition, rs136148 remained nominally associated with nondiabetic nephropathy in the FIND samples after adjustment for APOL1 G1 and G2 coding variants (p = 2.06 × 10−2). Based on these analyses and the Genovese et al. 7 and Tzur et al. 8 reports, variants in APOL1 are responsible for a significant portion of the increased susceptibility to nondiabetic nephropathy in AAs. Although the Genovese and Tzur reports identified APOL1 using SNPs from the 1000 Genomes Project, we were able to independently identify strong association with APOL1, relative to MYH9, using a pooled GWAS on the Illumina HumanHap550-Duo BeadChip. We note that APOL1 SNPs rs2239785 and rs136148 are not functional, whereas the G1 and G2 variants encode a non-conservative amino acid substitution and a 6–base pair deletion, respectively. G1 and G2 haplotypes are hypothesized to include causative kidney variants as well as protect from African sleeping sickness 28. Interestingly, the rs2239785 SNP is not in high linkage disequilibrium with either G1 or G2, (r2 values of 0.192 and 0.054 in AAs, respectively). These relatively low values, along with the results of the analyses conditional on the effects of G1 and G2, suggest that there may be additional functional variants in the region predisposing to kidney disease beyond the G1 and G2 variants.
There were several SNPs not on chromosome 22 that were modestly associated with nondiabetic nephropathy. The most associated was rs379489 in the CFH gene. CFH has reproducibly been associated with kidney disease, including IgA nephropathy 29, atypical hemolytic uremic syndrome (HUS) 30;31, glomerular C3 deposition with glomerulonephritis 24, and with rate of change in kidney function 32. It is possible that some of these AA cases on dialysis had IgA nephropathy, HUS, or glomerular C3 deposition but our participants did not have protocol or indication kidney biopsies to determine disease etiology. Alternatively, CFH may independently be associated with glomerulosclerosis.
Finally, modest association of a podocin (NPHS2) SNP was detected. NPHS2 is a podocyte-expressed gene associated with autosomal recessive glomerulosclerosis 13, microalbuminuria 27 and nephrotic syndrome 25. The same SNP in the NPHS2 gene was also the most strongly associated in the case-control interaction analysis with APOL1 risk variants. The mechanism by which podocin interacts with APOL1 to promote kidney disease is unknown at this time. The interaction between NPHS2 and APOL1, both clearly replicated and powerful effect nephropathy genes, is likely to be clinically relevant and potentially useful for screening in high risk populations. Interactions exist between Nphs2 and HIVAN1/HIVAN2 in the HIV-1 transgenic mouse model of collapsing FSGS 33, 34.
In conclusion, the APOL1 gene region is consistently and significantly associated with nondiabetic nephropathy among AAs, compared to controls without kidney disease. More modest association in the CFH gene appears relevant to nondiabetic kidney disease susceptibility in AAs, as well. Not only was the APOL1 association with nondiabetic nephropathy replicated in an almost equally large sample from FIND, we identified additional candidate genes associated with nondiabetic nephropathy and potentially interacting with APOL1, including NPHS2. Given the significance of the APOL1 association and the high incidence rates of nondiabetic nephropathy in the AA population, this study highlights the need for further investigation of these genes. It is possible that survival bias or selection bias can impact results involving prevalent patients on dialysis. Future studies should include more detailed genetic analyses of the candidate genes identified here, as well as functional studies focusing on the role of APOL1 in AAs with nondiabetic nephropathy and interactions between NPHS2 and APOL1.
A list of the members of the FIND Research Group follows (key: * Principal Investigator; ** Co-investigator; # Program Coordinator; § University of California, Davis; † University of California, Irvine; ‡ Study Chair). Genetic Analysis and Data Coordinating Center, Case Western Reserve University: SK Iyengar*, RC Elston**, KAB Goddard**, JM Olson**, S Ialacci#, J Fondran, A Horvath, R Igo, Jr., G Jun, K Kramp, J. Molineros, SRE Quade; Case Western Reserve University: JR Sedor*, J Schelling**, A Pickens#, L Humbert, L Getz-Fradley; Harbor-University of California Los Angeles Medical Center: S Adler*, E Ipp**, M Pahl**†, MF Seldin** §, S Snyder**, J Tayek**, E Hernandez#, J LaPage#, C Garcia, J Gonzalez, M Aguilar. Johns Hopkins University: M. Klag*, R. Parekh*, L Kao**, L Meoni**, T Whitehead, J Chester#; NIDDK, Phoenix, AZ: WC Knowler*, RL Hanson**, RG Nelson**, J Wolford**, L Jones#, R Juan, R Lovelace, C Luethe, LM Phillips, J Sewemaenewa, I Sili, B Waseta; University of California, Los Angeles: MF Saad*, SB Nicholas*, Y-D I Chen**, X Guo**, J Rotter**, K Taylor**, M Budgett, F Hariri#; University of New Mexico, Albuquerque: P Zager*, V Shah**, M Scavini**, A Bobelu#; University of Texas Health Science Center at San Antonio: H Abboud*, N Arar**, R Duggirala**, BS Kasinath**, F Thameem**, M Stern**; Wake-Forest University: BI Freedman*‡, DW Bowden**, CD Langefeld**, SC Satko**, SS Rich**, S Warren#, S Viverette, G Brooks, R Young, M Spainhour; Laboratory of Genomic Diversity, National Cancer Institute, Frederick, MD: C Winkler*, MW Smith**, M Thompson, R Hanson#, B Kessing; Minority Recruitment Centers: Loyola University: DJ Leehey*, G Barone#; University of Alabama at Birmingham: D Thornley-Brown*, C Jefferson#; University of Chicago: OF Kohn*, CS Brown#; NIDDK program office: JP Briggs, PL Kimmel, R Rasooly; External Advisory Committee: D Warnock (chair), L Cardon, R Chakraborty, GM Dunston, T Hostetter, SJ O’Brien (ad hoc), J Rioux, R Spielman.
We gratefully acknowledge the contributions of the Wake Forest participants and coordinators Joyce Byers, Carrie Smith, Mitzie Spainhour, Cassandra Bethea, and Sharon Warren. We further appreciate the contributions of FIND participants and physicians and CHOICE patients, staff, laboratory and physicians at Dialysis Clinic, Inc and Johns Hopkins University.
Support: This study was supported in part by NIH grants R01 DK 070941 and R01 DK 084149 (Dr Freedman), and R01 DK53591 (Dr Bostrom). Dr. Bostrom was supported by F32 DK080617 from the NIDDK. Computing resources were provided by the Wake Forest University Health Sciences Center for Public Health Genomics. This study was also supported by FIND grants U01DK57292, U01DK57329, U01DK057300, U01DK057298, U01DK057249, U01DK57295, U01DK070657, U01DK057303, U01DK070657, U01DK57304, and CHOICE study DK07024 from NIDDK and, in part, by the Intramural Research Program of the NIDDK. This project has been funded in whole or in part with federal funds from the NIH National Cancer Institute (NCI), under contract N01-CO-12400 and the Intramural Research Program of the NIH-NCI Center for Cancer Research. This work was also supported by the National Center for Research Resources for the General Clinical Research Center grants: Case Western Reserve University, M01-RR-000080; Wake Forest University, M01-RR-07122; Harbor–University of California, Los Angeles Medical Center, M01-RR-00425; College of Medicine, University of California, Irvine, M01-RR-00827-29; University of New Mexico, HSC M01-RR-00997; and Frederic C. Bartter, M01-RR-01346. The CHOICE study was supported in part by HS08365 from the Agency for Healthcare Research and Quality, Rockville, MD and HL62985 from the National Heart Lung and Blood Institute, Bethesda, MD. Genotyping was performed by the Center for Inherited Disease Research, which is fully funded through a federal contract from the NIH to Johns Hopkins University (N01-HG-65403).
Financial Disclosure: The authors declare that they have no other relevant financial interests.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.