PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
J Invest Dermatol. Author manuscript; available in PMC Dec 4, 2012.
Published in final edited form as:
PMCID: PMC3513759
NIHMSID: NIHMS422968
Fine-Mapping of Vitiligo Susceptibility Loci on Chromosomes 7 and 9 and Interactions with NLRP1 (NALP1)
Ying Jin,1 Sheri L. Riccardi,1 Katherine Gowan,1 Pamela R. Fain,1,2 and Richard A. Spritz1
1Human Medical Genetics Program, University of Colorado Denver, Anschutz Medical Campus, Aurora, Colorado, USA
2Barbara Davis Center for Childhood Diabetes, University of Colorado Denver, Anschutz Medical Campus, Aurora, Colorado, USA
Correspondence: Dr Richard A. Spritz, Human Medical Genetics Program, University of Colorado Denver, Anschutz Medical Campus, P.O. Box 6511, MS 8300, Aurora, CO 80045 USA. richard.spritz/at/ucdenver.edu
Generalized vitiligo is the most common pigmentation disorder, the result of autoimmune loss of melanocytes from the skin and hair, with a high frequency of other autoimmune diseases in vitiligo patients and their relatives. We previously reported the linkage signals on chromosomes 1, 7, and 17 in Caucasian families with generalized vitiligo and associated autoimmune diseases and identified the risk loci of chromosomes 17 and 1 as NLRP1 (NALP1) and FOXD3, respectively. Here, we describe fine-scale genetic association analyses in two independent series of Caucasian multiplex families, refining localization of the chromosome 7 locus and a locus on chromosome 9. Three susceptibility signals, represented by single-nucleotide polymorphisms (SNPs) rs6960920 in 7p13, rs734930 in 7q11, and rs4744411 in 9q22, were significantly associated with vitiligo and other autoimmune diseases. We also detected significant three-way interaction effects of chromosome 7 SNP rs6960920, chromosome 9 SNP rs4744411, and NLRP1 SNP rs6502867 on both the vitiligo phenotype and an expanded autoimmune disease phenotype, and significant three-way interaction effects of both chromosome 7 SNPs and NLRP1 SNP rs6502867 on the vitiligo phenotype. These support the validity of the chromosomes 7 and 9 linkage/association signals and underscore the utility of gene-gene interaction analysis in characterizing the genetic effects of candidate association signals.
Generalized vitiligo is an acquired, non-contagious disorder, in which progressive, patchy loss of pigmentation from the skin, overlying hair, and oral mucosa results from autoimmune loss of melanocytes from the involved areas (Nordlund et al., 2006). The most common pigmentation disorder, vitiligo, occurs in about 0.4% of European Caucasians (Howitz et al., 1977), with similar prevalence in other populations (Spritz, 2007, 2008). First-degree relatives of affected patients have ~6 to 7% risk of vitiligo, and concordance in monozygotic twins is 23% (Alkhateeb et al., 2003). About 25–30% of generalized vitiligo patients manifest at least one other autoimmune disease, particularly autoimmune thyroid disease (Graves’ disease and autoimmune hypothyroidism), rheumatoid arthritis, psoriasis, adult-onset autoimmune diabetes mellitus, pernicious anemia, Addison’s disease, and systemic lupus erythematosus; these same disorders also occur in increased frequency in patients’ first-degree relatives (Alkhateeb et al., 2003; Laberge et al., 2005). Thus, this complex of associated autoimmune diseases likely results from the interaction of multiple disease susceptibility variants (Majumder et al., 1993; Nath et al., 1994; Sun et al., 2006), some predisposing to a general autoimmune diathesis and others to specific autoimmune diseases.
A number of candidate genes have been reported as associated with generalized vitiligo, including loci in the major histocompatibility complex, ACE, CAT, CTLA4, COMT, ESR, GCH1, MBL, PTPN22, VDR, and others (Spritz, 2007, 2008). Most of these studies reported only marginally significant associations, often with no correction for multiple testing, and several were not replicated by subsequent studies (Spritz, 2007, 2008).
We previously reported significant linkage (logarithm of the odds (LODs) ≥3.0) of generalized vitiligo to loci on chromosomes 1, 7, 8, and 17 and suggestive linkage (LODs ≥2.0) to loci on chromosomes 9, 13, 19, and 22 in an analysis of 102 multiplex generalized vitiligo families. The linkage signals on chromosomes 1, 7, and 17 derived principally from the subset of 51 families segregating both vitiligo and additional vitiligo-associated autoimmune diseases, whereas the linkage signal on chromosome 8 was derived principally from the subset of families segregating only vitiligo (Spritz et al., 2004). Subsequently, we identified the chromosome 17 vitiligo-autoimmunity susceptibility locus as NLRP1 (NACHT leucine-rich-repeat protein) (Jin et al., 2007), encoding NALP1, a key regulator of the innate immune system, and the linkage signal on chromosome 1 derived primarily from a single large family (Alkhateeb et al., 2002), in which generalized vitiligo co-segregated with a promoter variant in FOXD3 (Alkhateeb et al., 2005).
The purpose of the current study was to refine localization of these putative vitiligo-autoimmune disease susceptibility loci. We first carried out improved genetic linkage analysis, increasing linkage support for the chromosomes 7 and 9 loci, whereas greatly reducing support for loci on chromosomes 13, 19, and 22. We then carried out fine-scale genetic association analyses through the chromosomes 7 and 9 linkage regions in the aforementioned 51 families that segregated both vitiligo and additional associated autoimmune diseases, detecting apparent association for 17 single-nucleotide polymorphisms (SNPs) from the chromosome 7 linkage region and 8 SNPs from the chromosome 9 linkage region. We then re-tested these SNPs in an independent series of 63 vitiligo-autoimmune disease families, with a joint analysis showing significant associations on both chromosomes after correction for multiple testing. We next investigated potential genetic interactions among these variants and with the known vitiligo susceptibility variants in NLRP1, detecting significant interactions that underscore the polygenic, complex nature of vitiligo and associated autoimmune diseases.
Genetic linkage analysis
We first reassessed the evidence for genetic linkage, filtering the previous microsatellite data (Spritz et al., 2004) using an improved error-checking routine implemented in Merlin 1.1.2 (Abecasis et al., 2002), and then re-analyzing linkage in the total 102 multiplex vitiligo families and the subset of 51 families that also segregated other autoimmune diseases using Allegro 1.2c (Gudbjartsson et al., 2000). With these enhanced analyses, the maximum multipoint LOD for chromosome 7 increased to 4.31 (P = 6.07 × 10−6) for the total 102 families and 4.01 (P = 1.52 × 10−6) for the 51 vitiligo-autoimmune disease families; the LOD maximum was located at 89.4 cM and the 1-LOD interval spanned B33.3 cM in 7q21. Similarly, a locus on chromosome 9, which previously showed only suggestive evidence for linkage in the total 102 vitiligo families (Spritz et al., 2004), now yielded a maximum multipoint LOD score of 2.37 (P = 2.24 10−4) for the total 102 families and 3.18 (P = 6.35 × 10−5) xfor the 51 vitiligoautoimmune disease families; the LOD maximum was at 88.1 cM and the 1-LOD interval spanned ~17.7 cM in 9q12-q22. In contrast, support for putative vitiligo susceptibility loci on chromosomes 8, 13, 19, and 22 was reduced substantially, below the threshold for suggestive linkage (LOD 1.9); accordingly, these three signals were not pursued further.
Family-based association studies
To refine localization of the chromosome 7 and 9 vitiligo-autoimmunity susceptibility loci, we next carried out family-based association analyses of high-density SNPs genotyped through the 1-LOD linkage intervals. We genotyped 333 members of the aforementioned 51 vitiligo-autoimmune disease families for 867 SNPs spanning the chromosome 7 linkage region and 304 SNPs spanning the chromosome 9 linkage region, respectively, capturing 38.6 and 27.2% of the common variation (minor allele frequency >0.1, r2>0.5) in these regions (NCBI Genome Build 35). Two SNPs from chromosome 7 and four from chromosome 9 deviated from the Hardy–Weinberg equilibrium (P<0.005) in unrelated founders and were excluded from further analyses. We used both the pedigree disequilibrium test (PDT) (Martin et al., 2000) and the family-based association test (FBAT) (Horvath et al., 2004) to assess the remaining markers for association with both vitiligo and with an expanded autoimmune disease phenotype that included family members with either vitiligo or any of the principal vitiligo-associated autoimmune diseases (Alkhateeb et al., 2003; Laberge et al., 2005; see Materials and Methods) as “affected.” SNPs showing nominal evidence of association (P<0.05) with vitiligo and/or the expanded autoimmunity phenotype, of which vitiligo is a subset, in both PDT and FBAT analyses (Figure 1, Tables S1, S2) were then genotyped in 323 family members from an independent series of 63 vitiligo-autoimmune disease families. Altogether, we tested 17 SNPs from the chromosome 7 linkage region and 8 SNPs from the chromosome 9 linkage region for association in the second series of 63 families, corresponding to 14 independent tests after allowing for marker-marker linkage disequilibrium (LD, Figure 2).
Figure 1
Figure 1
Association of generalized vitiligo and associated autoimmune diseases with chromosome 7 and 9 linkage region SNPs
Figure 2
Figure 2
LD patterns among the chromosome 7 and 9 linkage region SNPs
Of the 17 SNPs from the chromosome 7 linkage region showing both PDT and FBAT association with the vitiligo phenotype and/or with the expanded autoimmune disease phenotype in the first family series, four (rs6960920, rs2331071, rs2240100, and rs734930) SNPs also showed nominally significant association (P<0.05) in PDT analyses of both phenotypes and nominally significant or marginal association in FBAT analyses of the vitiligo phenotype in the second family series (Table 1, Table S1). The PDT P-values of these four SNPs in the combined sample of 114 families (series 1 and series 2 together) were significant for the vitiligo phenotype after the Bonferroni correction for the 14 independent tests of chromosome 7 and 9 SNPs (nominal P-valueo0.00357). SNP rs734930 showed the strongest evidence for association, with significant corrected P-values in PDT, FBAT, and conditional logistic regression analysis (CLRA) of the vitiligo phenotype (odds ratio (OR) 1.85), and PDT analysis of the expanded autoimmune disease phenotype (OR = 1.55). The other three chromosome 7 SNPs (rs6960920, rs2331071, and rs2240100) cluster in a block of strong LD (Figure 2).
Table 1
Table 1
P-values for allelic association of chromosome 7 variants with generalized vitiligo and an expanded autoimmune disease phenotype
To evaluate the relative effects of these four SNPs, we carried out stepwise CLRA of the case-pseudocontrol data set (Cordell and Clayton, 2002) derived from the combined 114 families. The effects of individual SNPs in the three-SNP cluster were indistinguishable by these analyses, and no haplotype explained the disease associations better than any individual SNPs (data not shown). However, including rs734930 significantly improved the fit of models that included any one of the three other SNPs; conversely, including any one of rs6960920, rs2331071, and rs2240100 significantly improved the fit of models that included rs734930 (Table 2). Thus, at least two independent chromosome 7 variants appear to contribute to genetic association with the vitiligo phenotype (Table 2), one tagged by rs734930 and the other tagged by the three-SNP cluster rs6960920-rs2331071-rs2240100.
Table 2
Table 2
Second locus effect tests of the chromosome 7 SNPs in a forward stepwise-regression procedure for the generalized vitiligo and expanded autoimmune disease phenotypes
Of the eight chromosome 9 SNPs showing both PDT and FBAT association with the vitiligo phenotype and/or the expanded autoimmune disease phenotype in the first family series, four (rs1350564, rs4744369, rs391784, and rs4744411) showed nominally significant or marginal association in FBAT analyses of both phenotypes and nominally significant association in PDT analyses of the expanded autoimmune disease phenotype in the second family series (Table 3, Table S2). The FBAT P-values of these four SNPs in the combined sample of 114 families were significant for both phenotypes after the Bonferroni correction for the 14 independent tests. SNP rs4744411 showed the strongest evidence for association with both phenotypes (OR = 2.03 for the vitiligo phenotype and OR = 2.05 for the expanded autoimmune disease phenotype).
Table 3
Table 3
P-values for allelic association of the chromosome 9 variants with generalized vitiligo and an expanded autoimmune disease phenotype
As all four of these chromosome 9 SNPs are clustered in a block of strong LD (Figure 2), we again used CLRA to evaluate the relative effects of individual SNPs. Including rs2584806 (vitiligo phenotype) or rs391784 (vitiligo phenotype and expanded autoimmune disease phenotype) significantly improved the fit of models that included rs4744369, but including rs4744369 did not significantly improve the fit of models that included either rs2584806 or rs391784 (Table 4). The individual effects of rs2584806, rs391784, and rs4744411 were indistinguishable by these analyses, and no SNP haplotype explained the disease associations better than did any individual SNPs (data not shown). Thus, the association of rs4744369 with disease is secondary to LD with the other SNPs.
Table 4
Table 4
Second locus effect tests of chromosome 9 SNPs in a forward stepwise-regression procedure for the generalized vitiligo and expanded autoimmune disease phenotypes
Genetic interaction analyses
We next explored potential genetic interactions among the two independent vitiligo susceptibility signals we detected on chromosome 7p and 7q (tagged by rs6960920 and rs734930, respectively), the signal on 9q (tagged by rs4744411), and the susceptibility variants of NLRP1 on chromosome 17p that we identified previously (tagged by rs6502867 and rs4790797) (Jin et al., 2007). We carried out CLRA in the combined case-pseudocontrol data set (Cordell and Clayton, 2002; Cordell et al., 2004a) of the five SNPs derived from the combined 114 families, using a three-stage interaction testing framework (Millstein et al., 2006). Analyses were restricted to two- and three-way interactions to avoid over-fitting the data, and tests at higher stage levels were conditional on effects that were already declared significant at lower stage levels to avoid multiple tests of the same effect. The significance threshold for an overall type 1 error rate of 0.05 was determined by the Bonferroni correction based on the number of stages and the number of tests within each stage. As in our previous report of disease associations of NLRP1 variants in these families (Jin et al., 2007), P-values for rs6502867 and rs4790797 met the significance threshold (P<0.0033) in stage 1 interaction testing framework analyses of both the vitiligo phenotype (Table 5, P=0.0022 and P = 0.0003, respectively) and the expanded =autoimmune disease phenotype (Table 6, P = 0.0005 and P=0.0001, respectively). Consistent with the results of univariate = CLRA (Tables 1 and and3),3), P-values for chromosome 7q SNP rs734930 and chromosome 9 SNP rs4744411 met the significance threshold in stage 1 interaction testing framework analyses (P<0.0033) of the vitiligo phenotype (Table 5, P = 0.0015 and P = 0.0011, respectively), and the P-value for chromosome 9 SNP rs4744411 on the expanded autoimmune disease phenotype was very close to significant (Table 6, P = 0.0036). Although P-values for the independent effect of chromosome 7p SNP rs6960920 were not significant foreither phenotype (P = 0.0141 for vitiligo, P = 0.1012 for the expanded autoimmune disease phenotype), there was suggestive evidence of a two-way interaction between rs6960920 and NLRP1 SNP rs6502867 (Table 5, P = 0.0019; nominal significance threshold P = 0.0017), as well as a significant three-way interaction between chromosome 7 SNPs rs6960920 and rs734930 and NLRP1 SNP rs6502867 (Table 5, P = 0.0009) for the vitiligo phenotype, and a significant three-way interaction between chromosome 7p SNP rs6960920, chromosome 9 SNP rs4744411, and NLRP1 SNP rs6502867 for both vitiligo (Table 5, P = 0.0012) and the expanded autoimmune disease phenotype (Table 6, P = 0.0006).
Table 5
Table 5
Wald tests of interactions for the generalized vitiligo phenotype using an interaction testing framework
Table 6
Table 6
Wald tests of interactions for the expanded autoimmune disease phenotype using an interaction testing framework
We have carried out a high-density SNP association study across regions of genetic linkage we previously detected on chromosome 7 and 9 for generalized vitiligo, an autoimmune disease of skin depigmentation. The family-based nature of the study makes the results robust against false-positive associations from cryptic population stratification. The linkage results suggested that loci in these regions might contribute to both generalized vitiligo and to a broader autoimmunity phenotype.
Three signals, tagged by SNPs rs6960920 in 7p13, rs734930 in 7q11, and rs4744411 in 9q22, were significantly associated with generalized vitiligo as well as with a broader autoimmunity phenotype that included vitiligo and other autoimmune diseases commonly associated with vitiligo. These three SNPs also showed significant interactions with NLRP1 SNP rs6502867, which we previously showed to be associated with vitiligo and other autoimmune diseases in these families (Jin et al., 2007). Both chromosome 7 signals yielded better P-values for the vitiligo phenotype than for the expanded autoimmunity phenotype, whereas the chromosome 9 signal yielded generally better P-values for the expanded autoimmunity phenotype. Nevertheless, a three-way interaction model that included SNP rs6960920 (7p13), rs4744411 (9q22), and rs6502867 (NLRP1) yielded a better P-value for the autoimmunity phenotype, suggesting that the 7p13 signal may contribute to broader autoimmune susceptibility.
The rs6960920-rs2331071-rs2240100 SNP cluster is located in a region of chromosome 7p13 that contains two genes, CAMK2B and NUDCD3. This SNP cluster is located 5′ to CAMK2B and 5 kb 3′ to NUDCD3, and rs2331071 and rs2240100 are located within NUDCD3. CAMK2B encodes an apparent member of the serine/threonine protein kinase and Ca(2+)/calmodulin-dependent protein kinase subfamilies that has an essential role in PAF-induced macrophage priming; CAMK2B is thus a tenable biological candidate genefor generalized vitiligo. NUDCD3 encodes a protein that functions to maintain the stability of dynein intermediate chain, and appears as an unlikely candidate gene for susceptibility to vitiligo and other autoimmune diseases. SNP rs734930 is located in 7q11.2, within an intron of AUTS2, a putative autism candidate gene of unknownfunction; and SNP rs4744411 is located in chromosome 9q22, within an intron of C9orf3, a predicted gene of unknown function. Further studies of these to identify the specific causal genes and causal disease susceptibility variants will elucidate new biological pathways that mediate susceptibility to generalized vitiligo and associated autoimmune diseases, facilitating the development of new approaches to the treatment or prevention of these disorders.
Subjects
Study subjects included 656 individuals from 114 Caucasian extended families with generalized vitiligo and additional autoimmune diseases from the United States and the United Kingdom. Families were selected based on the joint criteria of multiple family members affected with generalized vitiligo, and at least one family member affected with at least one of the other autoimmune diseases that are epidemiologically associated with generalized vitiligo (autoimmune thyroid disease, rheumatoid arthritis, adult-onset autoimmune diabetes mellitus, psoriasis, pernicious anemia, systemic lupus erythematosus, and Addison’s disease) (Alkhateeb et al., 2003; Laberge et al., 2005). We studied two family series. Series 1 included 51 extended families (333 individuals) that we had previously used to map a vitiligo-autoimmune disease locus to chromosome 1, 7, 8 and 17 (Spritz et al., 2004), and series 2 included 63 similar independent families (323 individuals). All families from series 1, and about half from series 2, are presented in Laberge et al., 2005.
All available relevant affected and unaffected family members completed a detailed questionnaire providing clinical history of about 50 autoimmune, autoinflammatory, and immune-related diseases. All data were reviewed by the study investigators, and most family members (affected and unaffected) were personally examined by one of the study staff. Clinical details of the 114 study families, including co-morbid autoimmune diseases, have been described previously (Jin et al., 2007). This study was approved bythe Colorado Multiple Institutional Review Board (COMIRB) and the South Thames Regional Multicentre Research Ethics Committee (MREC), and the Declaration of Helsinki protocols were followed. Written informed consent was obtained from all study participants.
Genotyping
DNA was prepared from peripheral blood or saliva using the Puregene Genomic DNA Purification Kit (Gentra Systems, Minneapolis, MN) or Oragene Collection Kit (Oragene; DNA Genotek, Ottawa, Ontario, Canada) respectively. We used the Illumina genotyping service (Illumina, San Diego, CA) to genotype the same 51 families (series 1) we previously used for genetic linkage analysis (Spritz et al., 2004), assaying 867 SNPs spanning the chromosome 7 linkage region (47 kb average spacing) and 304 SNPs spanning the chromosome 9 linkage region (42 kb average spacing). We next carried out a replication study in an independent series of 63 families (series 2) collected using the same criteria as series 1, genotyping the 17 SNPs from the chromosome 7 linkage region and the 8 SNPs from the chromosome 9 linkage region that yielded nominally significant association (P<0.05) by both PDT (Martin et al., 2000) and FBAT (Horvath et al., 2004) in series 1. Genotyping of the total 25 SNPs in series 2 was performed using the ABI Prism SNaPshot Multiplex kit on an ABI 3130XL DNA analyzer (Applied Biosystems, Foster City, CA) with HPLC-purified oligonucleotides (Operon, Huntsville, AL). To compare genotyping results obtained using the two platforms, we used the ABI system to re-genotype nine SNPs in the 333 individuals of series 1 and obtained a concordance rate of 99.8%.
Statistical analyses
Preliminary analyses
We assessed the inheritance of each marker in all families using PedCheck (O’Connell and Weeks, 1998) to test for inconsistencies owing to non-paternity, new mutations, or other errors.
Genetic linkage analysis
Analyses of our previous microsatellite data (Spritz et al., 2004) were carried out by applying the improved error-checking routine implemented in Merlin 1.0.0 (Abecasis et al., 2002) and then using Allegro 1.2c software (Gudbjartsson et al., 2000) to calculate LOD scores using an exponential model under the Spairs allele-sharing statistic.
Hardy-Weinberg equilibrium, LD, and disease association analyses
Testing for the Hardy-Weinberg equilibrium in founders and persons not in the line of descent, such as spouses, in all 114 families, and calculation of LD between NLRP1 region markers, was carried out using Haploview (Barrett et al., 2005) version 4.1. Calculation of the association of each marker with generalized vitiligo or with an expanded autoimmune phenotype that included any of generalized vitiligo, autoimmune thyroid disease, rheumatoid arthritis, adult-onset autoimmune diabetes mellitus, psoriasis, pernicious anemia, systemic lupus erythematosus, or Addison’s disease was carried out using PDT (Martin et al., 2000) and FBAT (Horvath et al., 2004). Haplotype-based transmission-disequilibrium statistics were calculated using FBAT.
To distinguish independent associations from secondary associations owing to LD, we applied CLRA (Cordell and Clayton, 2002) to the case-pseudocontrol data set derived from the combined 114 families. To allow for correlation between multiple affected individuals within a single pedigree, we clustered cases and matched pseudocontrols by pedigree and used robust the Huber–White variance estimators to assess the significance of each association (Cordell, 2004b). We assumed an additive genotypic effect for the high-risk allele of each locus. ORs were estimated using the coefficients from the regression equations for individual markers. As a test of the independent effect of a given locus conditioned on the effect of another locus, we compared the fit of a model containing both loci to a model containing only the conditioning locus (Cordell and Clayton, 2002). To test interactions among the chromosome 7 SNPs rs6960920 and rs734930, the chromosome 9 SNP rs4744411, and the NLRP1 SNPs rs6502867 and rs4790797, we carried out CLRA in the combined case-pseudocontrol data set (generated by combining the case-pseudocontrol data sets of the two chromosome 7 SNPs, the one chromosome 9 SNP, and the two NLRP1 SNPs derived from the combined 114 families) (Cordell and Clayton, 2002; Cordell et al., 2004a) using a three-stage interaction testing framework (Millstein et al., 2006) in order to test individual effects of signals represented by the five SNPs (stage 1), together with two-way interactions (stage 2) and with three-way interactions (stage 3). Analyses were carried out using STATA (http://www.stata.com), using the “geneassoc” routine (http://www-gene.cimr.cam.ac.uk/clayton/software/stata), and we assessed significance using a Wald test.
Supplementary Material
Supplement
ACKNOWLEDGMENTS
We thank the many families who participated in this study, and the UK Vitiligo Society, the US National Vitiligo Foundation, and Vitiligo Support International for their enthusiastic help in family ascertainment. We thank Anita Amadi-Myers and Paulene Holland for their invaluable assistance. This work was supported by Grants AR45584, AI46374, and DK57538 from the National Institutes of Health.
Abbreviations
CIconfidence interval
CLRAconditional logistic regression analysis
FBATfamily-based association test
LDlinkage disequilibrium
NLRP1NACHT leucine-rich-repeat protein
ORodds ratio
PDTpedigree disequilibrium test
SNPsingle-nucleotide polymorphism

Footnotes
CONFLICT OF INTEREST
The authors state no conflict of interest.
SUPPLEMENTARY MATERIAL
Supplementary material is linked to the online version of the paper at http://www.nature.com/jid
  • Abecasis GR, Cherny SS, Cookson WO, Cardon LR. MERLIN—rapid analysis of dense genetic maps using sparse gene flow trees. Nat Genet. 2002;30:97–101. [PubMed]
  • Alkhateeb A, Fain PR, Spritz RA. Candidate functional promoter variant in the FOXD3 melanoblast developmental regulator gene in autosomal dominant vitiligo. J Invest Dermatol. 2005;125:388–91. [PubMed]
  • Alkhateeb A, Fain PR, Thody A, Bennett DC, Spritz RA. Epidemiology of vitiligo and associated autoimmune diseases in Caucasian probands and their relatives. Pigment Cell Res. 2003;16:208–14. [PubMed]
  • Alkhateeb A, Stetler GL, Old W, Talbert J, Uhlhorn C, Taylor M, et al. Mapping of an autoimmunity susceptibility locus (AIS1) to chromosome 1p31.3-p32.2. Hum Mol Genet. 2002;11:661–7. [PubMed]
  • Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics. 2005;21:263–5. [PubMed]
  • Cordell HJ. Properties of case/pseudocontrol analysis for genetic association studies: effects of recombination, ascertainment, and multiple affected offspring. Genet Epidemiol. 2004b;26:186–205. [PubMed]
  • Cordell HJ, Barratt BJ, Clayton DG. Case/pseudocontrol analysis in genetic association studies: a unified framework for detection of genotype and haplotype associations, gene-gene and gene-environment interactions, and parent-of-origin effects. Genet Epidemiol. 2004a;26:167–85. [PubMed]
  • Cordell HJ, Clayton DG. A unified stepwise regression procedure for evaluating the relative effects of polymorphisms within a gene using case/control or family data: application to HLA in type 1 diabetes. Am J Hum Genet. 2002;70:124–41. [PubMed]
  • Gudbjartsson DF, Jonasson K, Frigge ML, Kong A. Allegro, a new computer program for multipoint linkage analysis. Nat Genet. 2000;25:12–3. [PubMed]
  • Horvath S, Xu X, Lake SL, Silverman EK, Weiss ST, Laird NM. Family-based tests for association haplotypes with general phenotype data: application to asthma genetics. Genet Epidemiol. 2004;26:61–9. [PubMed]
  • Howitz J, Brodthagen H, Schwartz M, Thompsen K. Prevalence of vitiligo: epidemiological survey of the Isle of Bornholm, Denmark. Arch Dermatol. 1977;113:47–52. [PubMed]
  • Jin Y, Mailloux CM, Gowan K, Riccardi SL, LaBerge G, Bennett DC, et al. NALP1 in vitiligo-associated multiple autoimmune disease. N Engl J Med. 2007;356:1216–25. [PubMed]
  • Laberge G, Mailloux CM, Gowan K, Holland P, Bennett DC, Fain PR, et al. Early disease onset and increased risk of other autoimmune diseases in familial generalized vitiligo. Pigment Cell Res. 2005;18:300–5. [PubMed]
  • Majumder PP, Nordlund JJ, Nath SK. Pattern of familial aggregation of vitiligo. Arch Dermatol. 1993;129:994–8. [PubMed]
  • Martin ER, Monks SA, Warren LL, Kaplan NL. A test for linkage and association in general pedigrees: the pedigree disequilibrium test. Am J Hum Genet. 2000;67:146–54. [PubMed]
  • Millstein J, Conti DV, Gilliland FD, Gauderman WJ. A testing framework for identifying susceptibility genes in the presence of epistasis. Am J Hum Genet. 2006;78:15–27. [PubMed]
  • Nath SK, Majumder PP, Nordlund JJ. Genetic epidemiology of vitiligo: multilocus recessivity cross-validated. Am J Hum Genet. 1994;55:981–90. [PubMed]
  • Nordlund JJ, Ortonne JP, Le Poole IC. Vitiligo vulgaris. In: Nordlund JJ, Boissy RE, Hearing VJ, King RA, Oetting WS, et al., editors. The pigmentary system. 2nd edn Blackwell Publishing; Malden, MA: 2006. pp. 551–98.
  • O’Connell JR, Weeks DE. PedCheck: a program for identification of genotype incompatibilities in linkage analysis. Am J Hum Genet. 1998;63:259–66. [PubMed]
  • Spritz RA. The genetics of generalized vitiligo and associated autoimmune diseases. Pigment Cell Res. 2007;20:271–8. [PubMed]
  • Spritz RA. The genetics of generalized vitiligo. Curr Dir Autoimmun. 2008;10:244–57. [PubMed]
  • Spritz RA, Gowan K, Bennett DC, Fain PR. Novel vitiligo susceptibility loci on chromosomes 7 (AIS2) and 8 (AIS3), confirmation of SLEV1 on chromosome 17, and their roles in an autoimmune diathesis. Am J Hum Genet. 2004;74:188–91. [PubMed]
  • Sun X, Xu A, Wei X, Ouyang J, Lu L, Chen M, et al. Genetic epidemiology of vitiligo: a study of 815 probands and their families from south China. Int J Dermatol. 2006;45:1176–81. [PubMed]