and Immunity (2006) 7, 169–178 & 2006 Nature Publishing Group All rights reserved 1466-4879/06 $30.00 www.nature.com/gene

ORIGINAL ARTICLE Variation in the type I cluster on 9p21 influences susceptibility to asthma and atopy

A Chan1,2, DL Newman1,3, AM Shon, DH Schneider, S Kuldanek and C Ober Department of Human Genetics, The University of Chicago, Chicago, IL, USA

A genome-wide screen for asthma and atopy susceptibility alleles conducted in the Hutterites, a founder population of European descent, reported evidence of linkage with a short tandem repeat polymorphism (STRP) within the type I interferon (IFN) gene cluster on 9p21. The goal of this study was to identify variation within the IFN gene cluster that influences susceptibility to asthma and atopic phenotypes. We screened approximately 25 kb of sequence, including the flanking sequence of all 15 functional genes and the single coding exon in 12, in Hutterites representing different IFNA-STRP genotypes. We identified 78 polymorphisms, and genotyped 40 of these (in 14 genes) in a large Hutterite pedigree. Modest associations (0.003oPo0.05) with asthma, bronchial hyper-responsiveness (BHR), and atopy were observed with individual variants or genes, spanning the entire 400 kb region. However, pairwise combinations of haplotypes between genes showed highly significant associations with different phenotypes (Po10À5) that were localized to specific pairs of genes or regions of this cluster. These results suggest that variation in multiple genes in the type I IFN cluster on 9p22 contribute to asthma and atopy susceptibility, and that not all genes contribute equally to all phenotypes. Genes and Immunity (2006) 7, 169–178. doi:10.1038/sj.gene.6364287; published online 26 January 2006

Keywords: asthma; atopy; type I

Introduction (P ¼ 4.9 Â 10À3). Both tests indicated that the 150 and 156 (bp) alleles were over-represented and the Asthma is a complex disease that affects nearly 300 154 and 158 bp alleles under-represented in affected million individuals worldwide.1 The rising incidence of individuals. The IFNA-STRP lies within the type I IFN asthma and atopic disorders over the past decades attests gene cluster, which spans approximately 400 kb on to the importance of environmental and lifestyle factors chromosome 9p22. The region consists of 15 functional in disease risk,2–4 although family and twin studies IFN genes, including 13 IFN-a genes, one IFN-o gene, support a strong genetic component to both.5,6 Allergic and one IFN-b gene, in addition to 11 IFN pseudogenes.14 sensitization to one or more allergens (atopy) is one of All IFN genes consist of a single exon and are highly the most important risk factors for asthma, as are family homologous to each other. On the whole, the coding history and early life environmental exposures.7–11 regions of the 13 IFN-a genes display greater than 80% To identify genes that might be involved in asthma nucleotide identity to each other, while the IFN-o gene and atopy pathogenesis, we conducted a genome-wide displays approximately 60% identity to any of the IFN-a screen in a founder population, the Hutterites.12 In this genes, and the IFN-b gene is about 40% identical to the earlier study, a short tandem repeat polymorphism rest of the cluster.15 (IFNA-STRP) on the short arm of showed The type I interferons are pleiotropic with evidence of linkage to asthma using the transmission important roles in inflammation, immunoregulation, and disequilibrium test (TDT) in a large Hutterite pedigree T cell responses,16 making them outstanding candidates (P ¼ 2.5 Â 10À5). More recently, using a case–control test,13 for asthma and atopy susceptibility genes. The potent this STRP was also associated with asthma antiviral properties of type I interferons have long been recognized,17,18 and viral infections, such as respiratory syncytial virus (RSV) and rhinovirus (RV), may be early Correspondence: Professor C Ober, Department of Human Genetics, 19–22 The University of Chicago, 920 E. 58th Street, Room CLSC 507C, life triggers of asthma. More recently, however, Chicago, IL 60636, USA. additional roles for interferons have been described, E-mail: [email protected] including skewing the adaptive immune system toward 1These authors contributed equally. a Th1 profile23 and the induction of apoptosis.24 2Current address: Department of Pediatrics, Weill Medical College, Furthermore, reduced levels of IFN-a in patients with Cornell University, New York, NY, USA. 3 the Th2-associated diseases asthma and atopy have been Current address: Department of Biological Sciences, Rochester 25,26 Institute of Technology, Rochester, NY, USA. reported, and treatment with IFN-a therapy improves 27–30 Received 30 August 2005; revised 29 November 2005; accepted 20 clinical symptoms in some patients. Thus, we December 2005; published online 26 January 2006 hypothesized that variation in one or more of these Type I interferons and asthma and atopy A Chan et al 170 genes could modulate expression levels or function Evaluation of phenotypes of the type I interferons and, as a result, influence Details of our protocols have been described pre- the development or course of asthma and atopic viously;12,34,35 the clinical characteristics of the study diseases. population are shown in Table 1. Briefly, 660 subjects The goal of this study was to identify polymorphisms were phenotyped. Individuals were considered to have in the 15 IFN genes and their flanking sequences, asthma if they (1) were bronchial hyper-responsive genotype selected single-nucleotide polymorphisms (X20% decrease in baseline forced expiratory volume

(SNPs) in a sample of 4700 Hutterites who are well in one second [FEV1] after inhalation of p25 mg characterized with respect to asthma and atopy pheno- methacholine/ml); (2) reported the presence of two or types, and test the hypothesis that variation in the type I three symptoms (wheeze, cough, shortness of breath); IFN gene cluster contributes to asthma and atopy and (3) had a doctor’s diagnosis of asthma. Eighty-three susceptibility. individuals met this definition of asthma, which was in general mild, as both oral steroid use and hospitaliza- tions were rare. Because during field trips to Hutterite Materials and methods colonies, we identified cases of undiagnosed asthma and because many Hutterites did not reliably recall past Subjects asthma symptoms (some did not understand the word The Hutterites are a founder population of European ‘wheeze’ because it is not in the German language), we descent. The 760 Hutterites in our study population also used a more relaxed diagnosis of asthma, referred to belong to a single 13-generation, 1623 member pedigree here as ‘possible asthma’ (N ¼ 236), that included and live on nine communal farms (colonies) in South additional individuals with bronchial hyper-responsive- Dakota.31 The Hutterites in our studies can be traced ness (BHR) (N ¼ 80) or at least two symptoms (N ¼ 73) or back to 62 Hutterite ancestors who lived in the early both (N ¼ 83). Four hundred twenty-four individuals 1700’s to early 1800’s.32 The relatively small number of who had neither symptoms nor BHR were considered as founders minimizes genetic heterogeneity in the popula- controls. tion, and their communal, farming lifestyle ensures that Atopy (allergic sensitization) was evaluated by mea- all members are exposed to a relatively uniform suring specific IgE response by skin prick test (SPT) to 14 environment. In particular, all meals are prepared and common allergens: house dust mites (HDM) (Dermato- eaten in a common kitchen and dining room, smoking is phagoides pteronyssinus, Dermatophagoides farinae), cock- prohibited, and exposures to allergens are relatively roaches (Blattela germanica, Periplaneta Americana), molds uniform among the colonies. Overall, the similar (Alternaria alternate, Cladosporium herbarum, Aspergillus environmental exposure should enhance the role of fumigatus), pollens (Lolium perenne, Ambrosia artemisifolia, genetic variation on disease risk in this population.33 Artemisia vulgaris, Quercus alba, Betula verrucosa), and We visited nine Hutterite colonies that were chosen animal danders (Felis domesticus, Canis familiaris). Be- based both on their location and to represent the cause few Hutterites were sensitized to animal danders different lines of Hutterite colony descent.32 Nearly all (Table 1), we did not analyze that as a separate category. Hutterites over the age of 5 who were present in the All tests included a negative (saline solution) and a colonies on the days of our visits participated in the positive (histamine) control. A positive ( þ ) SPT was study. The mean age of participants was 30.1 years (s.d. defined as equivalent to a 3 mm or greater wheal 18.2 years; range 6–89), and mean inbreeding coefficient diameter resulting from the corrected reaction area of the individuals in this sample is 0.034 (s.d. 0.015), (subtraction of saline control). Analyses were performed slightly greater than that of first cousins once removed.12 for five atopy categories: þ SPT to one or more allergen These studies were approved by the Institutional Review group (n ¼ 311), þ SPT to HDM (n ¼ 158), þ SPT to Boards at the University of Chicago and the University of cockroach (n ¼ 150), þ SPT to mold (n ¼ 75), and þ SPT South Dakota. to tree/pollen (n ¼ 158).

Table 1 Clinical characteristics of the Hutterites by asthma status

Unaffected Symptoms BHR Asthma

Sample size 424 73 80 83 Mean age (years) (s.d.) 30.4 (15.7) 32.1 (14.7) 25.6 (19.4) 23.4 (16.6) Sex ratio (M/F) 0.77 0.92 0.90 1.18 Mean log IgE (IU) (s.d.) 1.3 (0.7) 1.5 (0.8) 1.5 (0.8) 1.8 (0.8) % +SPTX1 allergen 42.1 52.1 50.6 48.1 % +SPTX2 allergens 31.2 39.7 39.1 40.3 % +SPT HDM 21.1 21.9 32.9 27.3 % +SPT cockroach 20.3 23.3 29.1 18.2 % +SPT molds 9.9 11.0 11.4 15.6 % +SPT pollens/trees 19.9 28.8 19.0 33.8 % +SPT animal danders 6.5 8.2 11.4 11.7

HDM (D. pteronyssinus, D. farinae), Animal Danders (F. domesticus, C. familiaris), Molds (A. alternata, C. herbarum, A. fumigatus), Cockroach (B. germanica, P. americana), Pollens/trees (L. perenne, A. artemisifolia, A. vulgaris, Q. alba, B. verrucosa).

Genes and Immunity Type I interferons and asthma and atopy A Chan et al 171 Polymorphism discovery stream regions of each gene to amplify both flanking We screened DNA from 12 Hutterites who were selected sequences and the exons. Lastly, primers were designed because they were heterozygous at the IFNA-STRP locus whenever possible to amplify regions containing known and were relatively distantly related within the Hutterite and validated polymorphisms. A description of all pedigree. Two siblings (individuals 8 and 12 in Figure 1) variants, primer sequences and PCR conditions is were included because they carried all four parental described in Table S1. at the IFN ‘locus’. Ten of the individuals in Screening for polymorphisms in the Hutterites was our screening set were unaffected controls, one (indivi- performed by denaturing high-performance liquid chro- dual 5 in Figure 1) was of unknown affection status, and matography (DHPLC).36 DHPLC detects polymorphisms one (individual 12 in Figure 1) had BHR. These 12 by separating heteroduplex from homoduplex DNA individuals carried all four of the IFNA-STRP alleles in fragments under partial denaturing conditions. Homo- different combinations. After these studies were com- duplex fragments show up as a single peak, while pleted, we re-sequenced the exons of three genes (IFNA1, heteroduplex fragments show up as a doublet. Prior to IFNA8, and IFNA21) in an additional 36 individuals (18 DHPLC, PCR products were denatured at 941C and then with asthma symptoms, BHR, or both and 18 unaffected) re-annealed by gradual cooling to 651C over 30 min to who were also relatively unrelated to each other and to allow the formation of homoduplex or heteroduplex the 12 individuals in the screening set, to confirm that DNA molecules. All samples were then compared by nonsynonymous (ns)SNPs that were present in dbSNP DHPLC using either a Varian Chromatography System but not in our 12-person screening set were, in fact, not (Walnut Creek, CA, USA) or a WAVEs DNA Fragment present in the Hutterites. Analysis System (Transgenomic, Inc., San Jose, CA, Our goal was to identify at least one polymorphism USA). each in the 50 and 30 flanking regions of the coding exon PCR products from samples with variant DHPLC of each gene with a minor allele frequency (MAF) 40.10 patterns were sequenced to determine the underlying and all polymorphisms within exons. We used this cutoff sequence variation, using BigDyes terminator chemistry because we had already detected linkage to this region (BigDyes Terminator v3.1 Cycle Sequencing Kit, Applied by the TDT. We assumed, therefore, that the suscept- Biosystems, Foster City, CA, USA). Products were ibility allele(s) had to be relatively common to generate analyzed on an ABI 3100 sequencer. The primers used to this signal. Because of the high homology between the for sequencing were the same as those used for PCR. The exons of the IFN genes, we designed pairs of primers sequences were analyzed using Sequenchert version that were complementary to the upstream and down- 3.1.1 (Gene Codes, Ann Arbor, MI, USA).

IFNA16IFNA16 IFNA-STRP IFNA10IFNA10 IFNA17 IFNB1 IFNB1 IFNW1IFNW1 IFNA7IFNA7 IFNA5 IFNA5 IFNA13 IFNA2 IFNA1 IFNA1 IFNA14 IFNA14 IFNA6 IFNA4 IFNA4 IFNA2 IFNA8 IFNA21IFNA21 IFNA8 tel cen 2 3 4 5 6 7 8 9 1 11 10 12 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 40 41 42 43 44 45 46 47 48 13 14 37 38 39 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 IFNA-STRP 1 156/158 2 156/158 3 156/158 4 156/158 5 156/158 6 156/158 7 154/156 8 154/156 9 150/154 10 150/156 11 150/156 12 150/156

Figure 1 Genotype patterns in 12 Hutterites representing all IFNA-STRP alleles (genotypes are shown in the column on the right). Green rectangles show the approximate location of each single exon gene (exons not drawn to scale). Seventy-eight polymorphisms (numbered 1–78 across the top) were identified in 12 individuals (numbered 1–12 in the left column). The genes in the region are shown across the top (pseudogenes are not shown). Homozygosity for the common allele are shown as yellow boxes, homozygosity for the minor allele as green boxes, and heterozygosity as red boxes. Exonic SNPs are denoted by white circles (synonymous SNPs), black circles (nonsynomymous SNPs) and black squares (frameshift SNPs). Arrows below the figure show the SNPs that were genotyped in the larger sample. The frequencies of the IFNA-STRP alleles in the population are: 150 bp ¼ 0.170, 154 bp ¼ 0.063, 156 bp ¼ 0.554, and 158 bp ¼ 0.213. Descriptions of each variant are presented in Table S1.

Genes and Immunity Type I interferons and asthma and atopy A Chan et al 172 Genotyping Results We genotyped in 760 Hutterites all of the polymorphisms Sequence variation and patterns of LD present in the 12-person screening set that were amen- able to one of the genotyping platforms used in our We screened approximately 25 kb of DNA, including laboratory (see Table S1 for a description of these flanking sequence in all 15 genes and the coding exon in methods). Because of the high sequence similarity 12 of these genes. Overall, we identified 72 SNPs and six between these genes, we could only genotype 40 in/dels, including at least one polymorphism in the polymorphisms (36 SNPs and 4 insertion/deletions flanking region of each gene (Table S1). Twelve exonic (in/dels)). However, nearly all of the untyped SNPs polymorphisms included four synonymous SNPs, seven identified in the Hutterites were in strong LD with a nsSNPs, and one frameshift (fs) SNP (Figure 1). Four of typed SNP in the screening set. The 760 Hutterites the exonic polymorphisms were previously reported: Asp57frameshift (171insA) and Ile184Arg in IFNA17, included the 660 individuals with phenotypes described - in Table 1. Genotypes were checked for Mendelian errors Cys20Stop in IFNA10, and Tyr51Tyr (153 C T) in IFNB1.39,40 using PedCheck.37 Hardy–Weinberg proportions were Four nsSNPs reported in dbSNP were not present in checked using a method designed for large pedigrees.38 the Hutterite screening set. These included one each in IFNA1 (rs2230050; freq. 0.031), IFNA8 (rs3739630; freq. Analysis of individual variants 0.055), IFNA21 (rs3750478; freq. 0.036), and IFNW1 We tested the individual variants for asthma and atopy (rs2230055; freq. 0.007). To confirm that these variants phenotypes using a case–control test designed specifi- were not present in the Hutterites, we re-sequenced in an cally for large pedigrees.13 This method takes into expanded screening set of 36 individuals, the exons of 4 account both the relatedness between individuals and the three genes with nsSNPs at 0.01 frequency in their inbreeding, and is more powerful than the TDT in dbSNP. None of the three nsSNPs (nor an additional IFNA21 the Hutterites.13 We corrected for multiple comparisons synonomous SNP in ; rs1053887, freq. 0.033) were using the Bonferroni correction for the number of genes present in the Hutterites (total of 96 chromosomes). Thus, consistent with a previous study in the Hutter- tested (N ¼ 14) and considered Po0.0036 (0.05/14) to be 41 significant. We did not correct for the number of ites, many SNPs that occur at frequencies o0.10 in the phenotypes (3 asthma and 5 atopy phenotypes) in any general population are absent in the Hutterites. dbSNP 4 analyses because these phenotypes are highly correlated also reports four common nsSNPs ( 0.10 frequency) in and the Bonferroni correction is overly conservative in IFNA4 (rs3750380, rs16938291, rs16938289, rs1062571), an such cases. exon that we were unable to specifically amplify. Three of these SNPs coincide with nucleotide differences between IFNA4 and the closely related IFNA17 gene, Haplotype analysis raising the possibility that they are not true SNPs. Haplotypes comprised of polymorphisms within each Because of this possibility and because we were not able gene were constructed manually by direct observation of to design primers that were specific to this gene, we did alleles segregating in families. Phase was assigned to not perform additional studies in IFNA4. each parental chromosome based on the genotype data in The 78 polymorphisms that were identified in the 12- families, which were often three generations. Haplotypes person screening set formed 22 unique genotype patterns were not constructed in situations in which phase could among the 12 subjects in our screening set (Figure 1). not be unambiguously determined. These gene-specific Even among individuals with the same IFNA-STRP haplotypes were tested for association using the case– genotypes (e.g. individuals 1–6, individuals 7–8, and control test to determine whether variation within each individuals 10–12 in Figure 1), the genotypes at many gene contributes to risk. In two genes (IFNA14 and IFN loci differed, particularly with respect to poly- morphisms proximal to IFNA2 and distal to IFNA16. IFNA10), we were only able to genotype one SNP. As a 2 result, we included these individual SNPs in the Consistent with this, there was little LD (r o0.30) haplotypes of their nearest neighboring genes (IFNA17 between genes at the most proximal (IFNA1, IFNA8) and IFNA16, respectively). For these individual analyses and most distal (IFNW1, IFNB1) ends of the region and a cluster of high LD in the middle of the region between of 12 ‘genes’, we used Po0.0041 (0.05/12) as the cutoff for significance at the 5% level. genes from IFNA2 to IFNA10, which included the IFNA- During the construction of haplotypes within each STRP (Figure 2). Of note is that the fsSNP and all but two gene, we kept track of the parental origin of each rare nsSNPs were present only in four individuals in the haplotype whenever possible. This allowed us to con- screening set, all of whom also carried the IFN-STRP struct haplotypes between all pairwise combinations of 150 bp allele (individuals 9–12; Figure 1). genes, which were also tested for association using the case–control test. The latter analyses were performed to Single variant analysis determine if combinations of genes contributed to risk; Although we originally aimed to genotype one variant here we used a critical P-value of o0.00055 to account for each in the upstream and downstream flanking se- the 91 pairwise comparisons (0.05/91). quences of each single exon gene, and all known variants in the exons, due to the high between genes we could not always achieve this goal. Analysis of linkage disequilibrium In particular, we were not able to specifically amplify the We assessed linkage disequilibrium (LD) between markers by exons of IFNA2, IFNA4, and IFNA7. In addition, estimating r2 using LDplotter (http://innateimmunity.net/ although two SNPs were identified in the flanking IIPGA2/Bioinformatics/pbtoldplotform). regions of IFNA7, we were unable to genotype either,

Genes and Immunity Type I interferons and asthma and atopy A Chan et al 173 IFNA16 IFNA10IFNA10 IFNA17 IFNB1 IFNB1 IFNW1IFNW1 IFNA7 IFNA5 IFNA5 IFNA2IFNA2 IFNA1 IFNA1 IFNA7 IFNA14 IFNA14 IFNA4IFNA4 IFNA13IFNA13 IFNA21IFNA21 IFNA6IFNA6 IFNA8 IFNA8 tel cen

1 4 5 6 7 8 13 14 16 19 20 21 27 28 29 32 34 35 36 38 39 40 41 45 47 50 52 56 57 58 59 60 62 63 66 68 7172 75 78

21.0658 21.0661 21.0677 21.0688 21.0690 21.1284 21.1330 21.1344 21.1546 21.1569 21.1719 21.1779 21.2013 21.2055 21.2056 Mb Distance21.2081 from 9p-ter 21.2161 21.2176 21.2177 21.2180 21.2193 21.2318 21.2937 21.2987 21.3400 21.3421 21.3424 21.3593 21.3594 21.3732 21.3739 21.3757 21.3974 21.3976 21.3988 21.4016 21.4019 21.4287 21.4321 21.4327 Figure 2 Pattern of LD (measured by r2) between SNPs across the type I IFN cluster on 9p21. Fifteen functional genes are shown across the top; pseudogenes are not shown. Green rectangles show the approximate location of each single exon gene (exons not drawn to scale). Arrows above each exon show the direction of transcription. Polymorphisms are numbered as in Figure 1 and Table 2. Mb distances from 9p-ter are shown on the diagonal.

and we were able to genotype only one SNP each in type ‘possible asthma’ (called ‘loose asthma’ in that IFNA14 and IFNA10. Lastly, we were unable to genotype study). The STRP remained the most significantly two synonymous and six nsSNPs (Figure 1, Table S1). associated marker with the ‘possible asthma’ phenotype SNP24 (Cys20Stop) was present in only one individual (9 (Po0.01), suggesting that there is likely additional in Figure 1) in the screening set, but this subject carried variation in this region that underlies the association one other unique SNP (SNP20), presumably on the same with the IFNA-STRP or that haplotypes or multiple haplotype, that was genotyped in the larger sample; and genes contribute to the association with asthma. None of SNP18 (Gln102Lys) was present in only one individual (6 the variants that tagged the haplotypes with nsSNPs or in Figure 1) who also carried another unique SNP (SNP7) the fsSNP were associated with asthma, although SNP45 that was genotyped in the larger sample. Four other showed modest association with BHR. untyped nsSNPs were present only in individuals 10–12 In contrast, variation in genes in the distal portion of (SNP30, Asp94His; SNP 31, Ile46Val; SNP 73, Val10Ala; the region was more strongly associated with atopic SNP74, Ala4163Gly), who were also the only carriers of phenotypes (Po0.01). SNP16 and SNP21 were associated ten of the genotyped SNPs (SNP24, SNP38, SNP39, with þ SPT to tree/pollen allergens and SNP4 was SNP40, SNP45, SNP50, SNP52, SNP54, SNP57, SNP62). associated with þ SPT to mold and to cockroach Thus, these typed SNPs served as surrogates for the allergens; only the association with SNP16 in IFNA21 untyped nsSNPs in association studies. Overall, we remained significant after correcting for multiple com- genotyped 760 individuals in the Hutterite pedigree for parisons. Other more modest associations with þ SPT to 40 polymorphisms in 14 of the 15 IFN genes, including mold and cockroach allergens were present at loci four exonic SNPs (Table 2). The mean inter-marker spanning from IFNB1 to IFNA6, but not including the distance was 9.4 kb between markers across the region IFNA-STRP, suggesting that different variation in this and 4.7 kb between markers in the genes. All markers region influences atopic and asthma phenotypes. The were in Hardy–Weinberg equilibrium (P40.05). four SNPs (SNP27, SNP38, SNP45, and SNP52) that The results of association studies with the individual tagged the haplotypes with the nsSNPs and the fsSNPs variants are shown in Table 3. There was modest showed modest association (0.054P40.01) with a association (Po0.05) with asthma- and atopy-related number of atopic phenotypes, but were not among the phenotypes across the entire region. No single marker most significant results. showed association with asthma (not shown), although the phenotypes ‘possible asthma’ and BHR showed some Haplotype analysis: within genes association with variants in genes spanning the cluster, To explore the possibility that unidentified variation on a but with the most significant associations with variants haplotype confers susceptibility, we first examined in the three genes just proximal to the IFNA-STRP. This is associations with haplotypes comprised of variants consistent with our earlier study, in which we reported within each of the genes (the single typed variant in an association between the IFNA-STRP and the pheno- IFNA10 was combined with IFNA16 variants and the

Genes and Immunity Type I interferons and asthma and atopy A Chan et al 174 Table 2 40 polymorphisms genotyped in 760 Hutterites

Gene Location SNP Polymorphism Coding change MAF Mb Percent with genotype

IFNB1 30-flanking 1 G/C 0.036 21.0658 98.7 30-flanking 4 1 bp in/del 0.195 21.0661 85.0 Exon 5 C/T Tyr51Tyr (TAC-TAT) 0.427 21.0677 86.5 50-flanking 6 G/C 0.339 21.0688 96.3 50-flanking 7 A/G 0.207 21.069 95.8 IFNW1 30-flanking 8 C/T 0.141 21.1284 95.4 50-flanking 13 A/G 0.1 21.133 95.0 50-flanking 14 A/G 0.101 21.1344 94.7 IFNA21 30-flanking 16 A/T 0.292 21.1546 96.6 50-flanking 19 C/T 0.419 21.1569 80.3 IFNA4 30-flanking 20 G/C 0.113 21.1719 88.0 50-flanking 21 A/G 0.306 21.1779 93.1 IFNA10 50-flanking 27 1 bp in/del 0.065 21.2013 87.5 IFNA16 30-flanking 28 G/T 0.055 21.2055 95.1 30-flanking 29 C/T 0.234 21.2056 95.3 50-flanking 32 G/T 0.375 21.2081 87.6 IFNA17 30-flanking 34 A/G 0.294 21.2161 88.8 Exon 35 T/G Ile184Arg (ATA-AGA) 0.176 21.2176 90.2 Exon 36 C/T Ser174Ser (TCT-TCC) 0.188 21.2177 84.9 Exon 38 1 bp in/del Asp57fs (GAC-AGAC) 0.064 21.218 84.4 50-flanking 39 A/G 0.065 21.2193 93.9 IFNA14 50-flanking 40 C/T 0.059 21.2318 92.9 IFNA5 30-flanking 41 A/G 0.163 21.2937 98.2 50-flanking 45 A/G 0.068 21.2987 87.2 IFNA6 30-flanking 47 A/G 0.158 21.34 84.4 50-flanking 50 A/G 0.064 21.3421 98.7 50-flanking 52 C/T 0.067 21.3424 88.5 IFNA13 50-flanking 56 A/C 0.064 21.3593 90.5 50-flanking 57 C/T 0.062 21.3594 92.8 IFNA2 30-flanking 58 A/G 0.057 21.3732 67.9 30-flanking 59 C/T 0.188 21.3739 95.4 50-flanking 60 6 bp in/del 0.041 21.3757 77.2 IFNA8 50-flanking 62 A/G 0.066 21.3974 98.4 50-flanking 63 A/G 0.463 21.3976 94.7 50-flanking 66 C/T 0.251 21.3988 97.4 30-flanking 68 A/G 0.463 21.4016 94.7 30-flanking 71 C/T 0.096 21.4019 98.3 IFNA1 50-flanking 72 A/G 0.383 21.4287 72.1 30-flanking 75 C/T 0.007 21.4321 98.0 30-flanking 78 A/G 0.172 21.4327 94.2

The minor allele at each SNP is shown on the right; the percentage of individuals in whom genotypes were successfully assigned is shown in the last column. MAF, minor allele frequency; Mb, distance from p-ter; fs, frameshift. rs numbers are shown in Table S1.

single typed variant in IFNA14 was combined with Haplotype analysis: between genes IFNA17 variants for a total of 12 ‘genic’ analyses). In this We next examined haplotype combinations between all analysis, three ‘genes’ in the distal end of the cluster pairs of genes to determine if variation in multiple genes were associated with atopic phenotypes: IFNA16/10 and interacts to influence risk for asthma or atopy. Here, IFNB1 with þ SPT to cockroach allergens (P ¼ 0.0031 and highly significant associations with both asthma and 0.0056, respectively) and IFNA21 with þ SPT to tree/ atopy phenotypes were observed, often between genes pollen allergens (P ¼ 0.0086). IFNA21 haplotypes were that show little or no LD (Figure 3). Eleven pairwise also more modestly associated with þ SPT to mold combinations of genes yielded 27 P-values o0.00055 allergens (P ¼ 0.012) and any allergen (P ¼ 0.028). None and, therefore, remained significant after adjusting for of the other genes were associated with any phenotype at multiple comparisons (Figure 3). The ‘genes’ most Po0.05. Only the associations with IFNA16/10 remained associated in the single gene analyses, IFNA16/10, significant after adjusting for multiple comparisons. The IFNA21, and IFNB1, were included in 7, 13, and 5 of IFNA16/10 haplotypes included variants in two genes, the significant pairwise associations, respectively. Addi- two untyped nsSNPs in IFNA16, and a variant (SNP27) tionally, IFNA1 was included in 11, IFNA8 in six, and that tagged the haplotype with three nsSNPs and one IFNW1 in five of the significant pairwise comparisons. fsSNP. Thus, similar to the single variant analysis, Together, at least one of these six ‘genes’ is included in IFNA21 and IFNB1 were associated with atopic pheno- all of the significant pairwise combinations. The five types and together with IFNA16/10 showed the most genes with nsSNPs (IFNA1, IFNA17, IFNA16, IFNA10, significant associations overall, and there were no IFNA21) were included in the most significant pairwise significant associations with asthma phenotypes. associations.

Genes and Immunity Type I interferons and asthma and atopy A Chan et al 175 Table 3 Results of single variant analyses in case–control tests

SNP/gene Asthma phenotypes +SPT to allergens

Possible asthma BHR Any Mold Cockroach Tree/pollen

4 IFNB1 0.006 0.005 6 IFNB1 0.048 0.015 7 IFNB1 0.033 13 IFNW1 0.026 14 IFNW1 0.025 16 IFNA21 0.038 0.023 0.002 21 IFNA4 0.009 27 IFNA10 0.045 0.014 0.027 28 IFNA16 0.017 32 IFNA16 0.045 0.042 35 IFNA17 0.028 38 IFNA17 0.025 0.044 0.047 45 IFNA5 0.049 0.037 0.016 0.034 52 IFNA6 0.033 IFNA-STRP 0.005 56 IFNA13 0.045 0.017 59 IFNA2 0.014 66 IFNA8 0.038 71 IFNA8 0.010

Uncorrected P-values are shown. Variants or phenotypes with at least one Po0.05 are shown; empty cells correspond to P40.05. None of the associations with asthma or +SPT to house dust mite had Po0.05 and are, therefore, not shown. See Materials and methods for description of phenotypes. IFNA14/17 IFNA16/10 IFNA21 IFNA13 IFNA8 IFNA4 IFNA1 IFNA6 IFNA5 IFNA2 IFNW IFNB (5) (6) (4) (2) (5) (2) (3) (3) (3) (2) (3) (2)

IFNA1 (3) IFNA8 M AS (5) IFNA2 M AS (3) IFNA13 M M AS (2) IFNA6 M B AS (3) IFNA5 M AS,B (2) IFNA14/17 AS,B (6) IFNA16/10 M,H C C AN,H AS,B (4) IFNA4 AN,M (2) H,C IFNA21 AN,M AN,H, (2) H,C T IFNW1 AN,M AN,H AN,H (3) H,C IFNB AN,C, AN,C (5) T

P < 0.00001 P < 0.0001 P < 0.001 P < 0.01 Figure 3 Results of association studies with all gene pairs. Haplotypes comprised of SNPs within each gene were considered as ‘alleles’, and then haplotypes comprised of all pairwise combinations of genes were constructed. The numbers of typed SNPs within each gene are shown in parentheses. (note that the single SNP in IFNA10 was combined with the SNPs in the adjacent IFNA16 gene and the single SNP in IFNA14 was combined with the SNPs in the adjacent IFNA17 gene). The three ‘genes’ that showed association individually with atopic phenotypes by the haplotype test at Po0.01 are shown on a shaded background. P-values corresponding to associations with possible asthma and BHR are shown in the upper right part of the figure and P-values corresponding to atopic phenotypes are shown in the lower left part of the figure. AS, possible asthma; B, bronchial hyper-responsiveness; AN, þ SPT to one or more allergens; M, þ SPT to mold allergens; H, þ SPT to house dust mite allergens; C, þ SPT to cockroach allergens; T, þ SPT to tree or pollen allergens.

Some combinations of genes appeared to be pheno- untyped nsSNPs, in combination with genes throughout type-specific, whereas others are associated with multi- most of the cluster is very specifically associated with ple phenotypes. For example, IFNA1, which includes two þ SPT to mold allergens, whereas combinations that

Genes and Immunity Type I interferons and asthma and atopy A Chan et al 176 include the four most proximal genes, IFNA4, IFNA21, combination with nearly all other genes with þ SPT to IFNAW, IFNAB, tend to be more associated with þ SPT mold allergens. Whereas þ SPT to house dust mite and to any allergen, house dust mite, cockroach, and/or tree/ cockroach (and trees/pollens to some extent) tended to pollen allergens. On the other hand, IFNA21 in combina- show associations with the same combinations of IFN tion with IFNA4 or IFNA16/10 is significantly associated genes, located mainly in the distal region of the cluster with possible asthma or BHR as well as with þ SPT to (IFNA4 to IFNB1), combinations of genes in the proximal any allergen and to HDM, cockroach allergens, or trees/ end (IFNA1 to IFNA5) showed associations with þ SPT pollens, and combinations of haplotypes in IFNB1 and to mold. This pattern of association suggests that the IFNA8 (genes located at opposite ends of the cluster) are genetics of the immune response to mold allergens significantly associated with possible asthma, and to differs from the genetics of the immune response to other þ SPT to any allergen, cockroach and tree/pollen. allergens. While we do not know the exact mechanism for these differences, we note that other studies in the Hutterites have also identified associations that were Discussion highly specific to mold allergens. For example, þ SPT to mold allergens was specifically associated with poly- We report here for the first time the patterns of variation morphisms in the integrin b3 gene, ITGB3, in the and LD in the type I IFN cluster on chromosome 9p21 Hutterites and three outbred populations45 and with and associations of asthma and atopy with variation in polymorphisms in the IL13 gene in the Hutterites.46 This IFN genes spanning this 400 kb region. We screened study further suggests that one or more variants in the B25 kb of DNA including the flanking regions of all 15 proximal end of the IFN cluster, which likely include functional genes and the single coding exon in 12 of these variation in IFNA1 or variation in LD with IFNA1, also genes, and identified 78 polymorphisms (72 SNPs, 6 in/ influences the immune response to inhaled mold aller- dels). The occurrence of one SNP every B330 bp is nearly gens. identical to the average rate reported in a survey of 106 Surprisingly, only IFNB1 in combination with IFNA8, genes.42 and IFNA21 in combination with IFNA16/10 or IFNA4, Curiously, despite the fact that these molecules have are associated with asthma or BHR, despite the fact that been implicated in multiple pathways that are critical in the original evidence for linkage in this region was with the development of immune-mediated diseases,16 little is an asthma phenotype. It is also noteworthy that no known about the specificity or regulation of individual associations were detected with a strict asthma pheno- genes encoding the type I IFNs. Although many of the type, but rather with a fairly loose definition of asthma functions of IFN-a and IFN-b overlap, responses to that included individuals with just symptoms or just treatments with each indicate that differences exist. For BHR or both, as in our previous study.12 Although it is example, IFN-b is used for the treatment of multiple likely that some Hutterites with BHR who did not recall sclerosis (MS) whereas IFN-a is more effective for symptoms and some of the symptomatic Hutterites who treatment of chronic viral hepatitis, chronic myelogenous did not have a 20% fall in baseline FEV1 on the day of our leukemia, ulcerative colitis, and steroid-resistant asth- testing do, in fact, have asthma, it is interesting that the ma.16 The side effects of each treatment also differ, strongest associations were with a fairly unconventional further suggesting functional specificities. In particular, definition of asthma. Moreover, these same pairs of genes IFN-a, but not IFN-b, therapy has been associated with were also associated with atopy phenotypes. Thus, we the development of acute arthritis or pancreatitis, suggest that these ‘asthma’ phenotypes are merely elevated levels of circulating complement, and glomer- expressions of a general immune dysregulation that is ulonephritis. Moreover, the regulation and timing of common to asthma/allergy symptoms and BHR, and expression of these genes may also impact their that genetic variation in the IFN genes influence this biological effects. In animal and cell culture models, the general immune response. effects of type I IFNs are both concentration- and Overall, these results implicate for the first time development-dependent. For example, IFN-a can induce genetic variation in the IFN gene cluster with Th2- type I diabetes in transgenic mice,16 but spontaneous mediated diseases, whereas prior studies of a limited diabetes in non-obese diabetic mice was inhibited by number of IFN variants have reported associations with IFN-a in a dose-dependent fashion.43 The addition of Th1-mediated diseases, MS40 and sarcoidosis.47 Although IFN-b to T cell–dendritic cell (DC) co-cultures during DC we cannot exclude the possibility that some of these maturation induced a type II IFN (IFN-g) response, but associations are due to typed or untyped coding addition of IFN-b to the primary co-culture shifted the variation, we note that the coding variation in the cells toward an IL-10 response and inhibited IFN-g Hutterites appears to be confined to one or a few production.44 These data suggest that there is tight haplotypes. Markers that tag these haplotypes, as well as regulation of expression of the type I IFN genes and the typed Asp57fs and Ile184Arg variants in IFNA17, that functional differences exist between IFN-b and IFN- were not strongly associated with any phenotypes in the a. single variant analysis, although most were infrequent The results of our study further suggest that the type I (B0.06) in the Hutterites and, as a result, our power to IFN genes may also differ with respect to regulation, detect associations with these SNPs was limited. The function, and impact on disease risk. Only four of the 14 exception to this is the variant tagging the haplotype genes studied were independently associated with risk with the Cys20Stop variant in IFNA10. This tag SNP for atopy (IFNA16/10, IFNA21, IFNB1). Other genes were (SNP20) occurs at frequency 0.11 but was not associated associated with risk but nearly always only in combina- with asthma or atopy in the single variant analysis, tion with one of the above genes. The notable exception suggesting that this nonsense mutation does not by itself to this is the highly specific association of IFNA1 in account for the observed associations. Taken together,

Genes and Immunity Type I interferons and asthma and atopy A Chan et al 177 our results do not provide evidence for coding variation 3 Beasley R. The burden of asthma with specific reference to the per se accounting for the associations with atopy or United States. J Allergy Clin Immunol 2002; 109: S482–S489. asthma, although this possibility cannot be ruled out. 4 Bach JF. The effect of infections on susceptibility to auto- On the other hand, variation in noncoding sequences immune and allergic diseases. N Engl J Med 2002; 347: 911–920. may influence the differential regulation and expression 5 Skadhauge LR, Christensen K, Kyvik KO, Sigsgaard T. Genetic levels of the IFN-a and b products. In fact, a recent study and environmental influence on asthma: a population-based of two coding variants in IFN genes in Japanese study of 11 688 Danish twin pairs. Eur Respir J 1999; 13: 8–14. sarcoidosis patients reported an association with a 6 Palmer LJ, Burton PR, James AL, Musk AW, Cookson WO. Familial aggregation and heritability of asthma-associated haplotype comprised of the IFNA10 Cys60Stop and quantitative traits in a population-based sample of nuclear IFNA17 Ile184Arg SNPs (SNP24 and SNP35, respectively, families. Eur J Hum Genet 2000; 8: 853–860. 48 in our study). Despite the fact that the association was 7 Burrows B, Lebowitz MD, Barbee RA. Respiratory disorders with two coding polymorphisms, cells from individuals and allergy skin-test reactions. Ann Intern Med 1976; 84: with different haplotype combinations showed different 134–139. expression levels of IFN-a , with the high IFN-a- 8 Martinez FD, Wright AL, Taussig LM, Holberg CJ, Halonen M, producing haplotype associated with increased risk for Morgan WJ. Asthma and wheezing in the first six years of life. sarcoidosis, consistent with the disease resulting from a The Group Health Medical Associates. N Engl J Med 1995; 332: Th1-mediated mechanism.47 This result was interpreted 133–138. as evidence for the presence of additional variation on 9 Peden DB. Development of atopy and asthma: candidate environmental influences and important periods of exposure. the associated haplotype that influences transcription or Environ Health Perspect 2000; 108 (Suppl 3): 475–482. expression of IFN genes. Our study, which implicates 10 Illi S, von Mutius E, Lau S, Nickle R, Niggemann B, noncoding variation in the IFN gene cluster in risk for Sommerfeld C et al. The pattern of atopic sensitization is Th2-mediated diseases, is consistent with the suggestion associated with the development of asthma in childhood. J that disease-associated variation affects expression levels Allergy Clin Immunol 2001; 108: 709–714. of one or more genes. 11 Lemanske RF. The childhood origins of asthma (COAST) The fact that the most significant associations were study. Pediatr Allergy Immunol 2002; 15: 1–6. with combinations of genes, often between genes that 12 Ober C, Tsalenko A, Parry R, Cox NJ. A second-generation show little or no LD, further suggests that these genes genomewide screen for asthma-susceptibility alleles in a may interact to influence disease course or pathogenesis. founder population. Am J Hum Genet 2000; 67: 1154–1162. Although we were unable to attribute differences in risk 13 Bourgain C, Hoffjan S, Nicolae R, Newman D, Steiner L, Walker K et al. Novel case–control test in a founder population of developing asthma or atopy to specific variants, identifies P-selectin as an atopy-susceptibility locus. Am J Hum polymorphisms in five genes, IFNA1, IFNA16, IFNA10, Genet 2003; 73: 612–626. IFNA21 and IFNB1, accounted for nearly all of the 14 Diaz MO. The human type I interferon gene cluster. Semin significant associations with atopy. IFNA8, IFNA2, and Virol 1995; 6: 143–149. IFNA4 in combination with one or more of those five 15 LaFleur DW, Nardelli B, Tsareva T, Mather D, Feng P, genes also showed significant associations with asthma. Semenuk M et al. Interferon-kappa, a novel type I interferon It is possible that variation in all eight genes influences expressed in human keratinocytes. J Biol Chem 2001; 276: IFN expression or function; comprehensive studies of 39765–39771. variation in these genes in additional patient populations 16 Tilg H, Kaser A. Type I interferons and their therapeutic role could help to refine the associations to fewer genes or in Th2-regulated inflammatory disorders. Expert Opin Biol Ther 2004; 4: 469–481. variants. Finally, functional studies of the associated 17 Foster GR, Finter NB. Are all type I human interferons noncoding variation could reveal which variants directly equivalent? J Viral Hepat 1998; 5: 143–152. influence expression levels and disease risk. Regardless, 18 Stark GR, Kerr IM, Williams BR, Silverman RH, Schreiber RD. our study identifies a new family of candidate genes for How cells respond to interferons. Annu Rev Biochem 1998; 67: asthma/atopy association studies, for studies of the 227–264. relationship between viral infections and the develop- 19 Sigurs N, Bjarnason R, Sigurbergsson F, Kjellman B. Respira- ment or exacerbation of asthma,49,50 and for pharmaco- tory syncytial virus bronchiolitis in infancy is an important genetic studies of the differential response to IFN-a risk factor for asthma and allergy at age 7. Am J Respir Crit Care therapy in corticosteroid-resistant asthma and other Med 2000; 161: 1501–1507. atopic diseases.27–30 20 Hall CB. Respiratory syncytial virus and parainfluenza virus. N Engl J Med 2001; 344: 1917–1928. 21 Psarras S, Papadopoulos NG, Johnston SL. Pathogenesis of respiratory syncytial virus bronchiolitis-related wheezing. Acknowledgements Paediatr Respir Rev 2004; 5 (Suppl A): S179–S184. 22 Gern JE, Martin MS, Anklam KA, Shen K, Roberg KA, We acknowledge Harvey Dytch, Emma Thompson, and Carlson-Dakes KT et al. Relationships among specific viral D’Anne Duncan for technical assistance. This work was pathogens, virus-induced -8, and respiratory supported by NIH Grants HL56399 and HL66533. symptoms in infancy. Pediatr Allergy Immunol 2002; 13: 386– 393. 23 Tilg H. New insights into the mechanisms of : References an immunoregulatory and anti-inflammatory . Gas- troenterology 1997; 112: 1017–1021. 1 Masoli M, Fabian D, Holt S, Beasley R. The global burden of 24 Kaser A, Nagata S, Tilg H. Interferon alpha augments asthma: executive summary of the GINA Dissemination activation-induced T cell death by upregulation of Fas Committee Report. Allergy 2004; 59: 469–478. (CD95/APO-1) and expression. Cytokine 1999; 11: 2 Burr ML, Butland BK, King S, Vaughan-Williams E. Changes 736–743. in asthma prevalence: two surveys 15 years apart. Arch Dis 25 Bufe A, Gehlhar K, Grage-Griebenow E, Ernst M. Atopic Child 1989; 64: 1452–1456. phenotype in children is associated with decreased virus-

Genes and Immunity Type I interferons and asthma and atopy A Chan et al 178 induced interferon-alpha release. Int Arch Allergy Immunol 39 Golovleva I, Kandefer-Szerszen M, Beckman L, Lundgren E. 2002; 127: 82–88. Polymorphism in the interferon-alpha gene family. Am J Hum 26 Dolen JG, Mathur A. Undetectable interferon-alpha serum Genet 1996; 59: 570–578. levels in a patient with atopic dermatitis. J Interferon Cytokine 40 Miterski B, Jaeckel S, Epplen JT, Pohlau D, Hardt C. The Res 1995; 15: 973–975. interferon gene cluster: a candidate region for MS predisposi- 27 Gratzl S, Palca A, Schmitz M, Simon HU. Treatment with IFN- tion? Multiple Sclerosis Study Group. Genes Immun 1999; 1: alpha in corticosteroid-unresponsive asthma. J Allergy Clin 37–44. Immunol 2000; 105: 1035–1036. 41 Newman DL, Hoffjan S, Bourgain C, Abney M, Nicolae RI, 28 Kimata H, Akiyama Y, Kubota M, Furusho K. Interferon-alpha treatment for severe atopic dermatitis. Allergy 1995; 50: 837– Profits ET et al. Are common disease susceptibility alleles the 840. same in outbred and founder populations? Eur J Hum Genet 29 Pung YH, Vetro SW, Bellanti JA. Use of interferons in atopic 2004; 12: 584–590. (IgE-mediated) diseases. Ann Allergy 1993; 71: 234–238. 42 Cargill M, Altshuler D, Ireland J, Sklar P, Ardlie K, Patil N et al. 30 Simon HU, Seelbach H, Ehmann R, Schmitz M. Clinical and Characterization of single-nucleotide polymorphisms in cod- immunological effects of low-dose IFN-alpha treatment in ing regions of human genes. Nat Genet 1999; 22: 231–238. patients with corticosteroid-resistant asthma. Allergy 2003; 58: 43 Sobel DO, Ahvazi B. Alpha-interferon inhibits the develop- 1250–1255. ment of diabetes in NOD mice. Diabetes 1998; 47: 1867–1872. 31 Abney M, McPeek MS, Ober C. Estimation of variance 44 Nagai T, Devergne O, Mueller TF, Perkins DL, van Seventer components of quantitative traits in inbred populations. Am JM, van Seventer GA. Timing of IFN-beta exposure during J Hum Genet 2000; 66: 629–650. human dendritic cell maturation and naive Th cell stimulation 32 Ober C, Weitkamp LR, Cox N, Dytch H, Kostyu D, Elias S. has contrasting effects on Th1 subset generation: a role for HLA and mate choice in humans. Am J Hum Genet 1997; 61: 497–504. IFN-beta-mediated regulation of IL-12 family cytokines and 33 Ober C, Abney M, McPeek MS. The genetic dissection of IL-18 in naive Th cell differentiation. J Immunol 2003; 171: complex traits in a founder population. Am J Hum Genet 2001; 5233–5243. 69: 1068–1079. 45 Weiss LA, Lester LA, Gern JE, Wolf RL, Parry R, Lemanske RF 34 Ober C, Tsalenko A, Willadsen S, Newman D, Daniel R, Wu X et al. Variation in ITGB3 is associated with asthma and et al. Genome-wide screen for atopy susceptibility alleles in sensitization to mold allergen in four populations. Am J Respir the Hutterites. Clin Exp Allergy 1999; 29 (Suppl 4): 11–15. Crit Care Med 2005; 172: 67–73. 35 Ober C, Cox NJ, Abney M, Di Rienzo A, Lander ES, 46 Donfack J, Schneider DH, Tan Z, Kurz T, Dubchak I, Frazer Changyaleket B et al. Genome-wide search for asthma KA, Ober C. Variation in conserved non-coding sequences on susceptibility loci in a founder population. The Collaborative chromosome 5q and susceptibility to asthma and atopy. Resp Study on the Genetics of Asthma. Hum Mol Genet 1998; 7: Res 2005; 6: 145. 1393–1398. 47 Newman LS, Rose CS, Maier LA. Sarcoidosis. N Engl J Med 36 Oefner P, Underhill P. Comparative DNA sequencing by 1997; 336: 1224–1234. denaturing high-performance liquid chromatography 48 Akahoshi M, Ishihara M, Remus N, Uno K, Miyake K, Hirota (DHPLC). Am J Hum Genet 1995; 57 (Suppl): A266. 37 O’Connell JR, Weeks DE. PedCheck: a program for identifica- T et al. Association between IFNA genotype and the risk of tion of genotype incompatibilities in linkage analysis. Am J sarcoidosis. Hum Genet 2004; 114: 503–509. Hum Genet 1998; 63: 259–266. 49 Martinez FD, Helms PJ. Types of asthma and wheezing. Eur 38 Bourgain C, Abney M, Schneider D, Ober C, McPeek MS. Respir J Suppl 1998; 27: 3s–8s. Testing for Hardy–Weinberg equilibrium in samples with 50 Gern JE. Viral respiratory infection and the link to asthma. related individuals. Genetics 2004; 168: 2349–2361. Pediatr Infect Dis J 2004; 23: S78–S86.

Supplementary Information accompanies the paper on Genes and Immunity’s website (http://www.nature.com/gene)

Genes and Immunity