PMC:4129407 / 5548-13811 JSONTXT

{"project":"2_test","denotations":[{"id":"25087609-19169254-2044411","span":{"begin":563,"end":565},"obj":"19169254"},{"id":"25087609-20953188-2044411","span":{"begin":563,"end":565},"obj":"20953188"},{"id":"25087609-23143594-2044411","span":{"begin":563,"end":565},"obj":"23143594"},{"id":"25087609-19169254-2044412","span":{"begin":683,"end":685},"obj":"19169254"},{"id":"25087609-20953188-2044412","span":{"begin":683,"end":685},"obj":"20953188"},{"id":"25087609-23143594-2044412","span":{"begin":683,"end":685},"obj":"23143594"},{"id":"25087609-16871531-2044413","span":{"begin":1233,"end":1235},"obj":"16871531"},{"id":"25087609-23762245-2044414","span":{"begin":1530,"end":1532},"obj":"23762245"},{"id":"25087609-23762245-2044415","span":{"begin":2032,"end":2034},"obj":"23762245"},{"id":"25087609-24656864-2044415","span":{"begin":2032,"end":2034},"obj":"24656864"},{"id":"25087609-17130525-2044415","span":{"begin":2032,"end":2034},"obj":"17130525"},{"id":"25087609-23080122-2044416","span":{"begin":2134,"end":2136},"obj":"23080122"},{"id":"25087609-23762245-2044417","span":{"begin":2363,"end":2365},"obj":"23762245"},{"id":"25087609-23762245-2044418","span":{"begin":2523,"end":2525},"obj":"23762245"},{"id":"25087609-23080122-2044419","span":{"begin":2979,"end":2981},"obj":"23080122"},{"id":"25087609-23762245-2044420","span":{"begin":3193,"end":3195},"obj":"23762245"},{"id":"25087609-23762245-2044421","span":{"begin":3703,"end":3705},"obj":"23762245"},{"id":"25087609-24656864-2044421","span":{"begin":3703,"end":3705},"obj":"24656864"},{"id":"25087609-11422182-2044422","span":{"begin":7207,"end":7208},"obj":"11422182"},{"id":"25087609-24656864-2044423","span":{"begin":7411,"end":7413},"obj":"24656864"}],"text":"Material and Methods\n\nSamples\nWe used data from 9,247 PsV-affected individuals and 13,589 control individuals obtained from six case-control PsV data sets, including four GWASs (the Collaborative Association Study of Psoriasis [CASP] and the Genizon, Kiel, and PsA GWASs), a targeted deep follow-up study of CASP (the CASP-DFU), and one Immunochip-based data set of 3,723 affected and 7,595 control subjects (the Psoriasis Association Genetics Extension [PAGE] study), for a total of 9,247 affected and 13,589 control individuals (Table S1, available online).7,9,12 Genotype data of the studies were generated and stringently quality-control (QC) filtered as described elsewhere,7,9,12 and all samples were confirmed to be unrelated individuals of European ancestry according to self-reported ethnicity and results of principal-component (PC) analysis. All participating individuals provided written informed consent and were recruited according to the protocols approved by the institutional review board of each institution.\n\nPhenotype Classification\nAll PsV-affected individuals were diagnosed by a dermatologist. Diagnosis of PsA was confirmed by a rheumatologist according to the Classification Criteria for Psoriatic Arthritis.30 Individuals who had had PsV for 10 or more years but no signs of PsA were classified as having PsC. Our data set included 3,038 PsA subjects, 3,098 PsC subjects, and 3,111 subjects of unknown PsA or PsC status (Table S1).\n\nStatistical Analysis\n\nHLA Imputation\nFor each data set, we used SNP2HLA24 to extract SNP genotypes located in the MHC region to impute classical two- and four-digit HLA alleles of and amino acid polymorphisms encoded by the eight class I and class II HLA genes (HLA-A, HLA-B, HLA-C, HLA-DRB1, HLA-DQA1, HLA-DQB1, HLA-DPA1, and HLA-DPB1). We conducted HLA imputation for each data set separately by using HLA and SNP genotypes from the Type 1 Diabetes Genetics Consortium (T1DGC; n = 5,225), which has demonstrated a high imputation accuracy for classical HLA alleles,24,26,31 as a reference panel. We obtained information on HLA-gene polymorphisms from the IMGT/HLA Database.32 Amino acid sequences encoded by the imputed HLA genes are indicated in Figure S1. For HLA amino acid positions, we indicate the start codon of the mature HLA protein as position 1, and we label the codon 5′ to this site as −1.24 SNP2HLA checks concordance of allele strands of the A/T or G/C SNPs between the data set and the reference panel on the basis of allele-frequency comparison.24 We applied postimputation QC criteria of MAF \u003e 0.1% for the association analysis.\n\nMICA Imputation\nTo expand our HLA imputation protocol into HLA-like genes, we constructed a reference panel for imputation of MICA variants. We obtained classical four-digit MICA alleles for the subjects from a subset of the PsA data set (n = 1,046). These samples were not selected in any particular way. We obtained MICA amino acid sequences from the IMGT/HLA Database32 and the encoded MICA amino acid polymorphisms of the subjects, as well as the genotypes of MICA classical alleles and the genotyped SNPs in the MHC region. Using the constructed MICA reference panel and SNP2HLA,24 we imputed MICA variants for the other data-set collections. Imputed genotypes of the MICA alleles and MICA amino acid polymorphisms were extracted and merged into those obtained from HLA imputation mentioned in the previous section. We empirically assessed the accuracy of imputing MICA variants by additionally genotyping MICA in a subset of the subjects from the PAGE Immunochip data set (n = 104) and comparing concordances of the imputed and genotyped classical MICA variants as described elsewhere.24,26\n\nStatistical Framework for Association Analysis\nWe used the following analyses to test associations between HLA variants and risk of four binary phenotypes: (1) overall analysis of PsV susceptibility (PsV-affected versus control individuals), (2) stratified analysis of PsA susceptibility (PsA-affected versus control individuals), (3) stratified analysis of PsC susceptibility (PsC-affected versus control individuals), and (4) intra-PsV analysis directly comparing PsA to PsC (PsA-affected versus PsC-affected individuals). For each phenotype, we assessed variant risk with a logistic-regression model assuming additive effects of the allele dosages in the log-odds scale and their fixed effects among the data-set collections. We defined HLA variants to include biallelic SNPs in the MHC region, two- and four-digit biallelic classical HLA or MICA alleles, biallelic HLA or MICA amino acid polymorphisms for respective residues, and multiallelic HLA or MICA amino acid polymorphisms for respective positions. To account for potential population-based and data-set-specific confounding factors, we included the top ten PCs and an indicator variable for each data set as covariates. For HLA variants with m alleles (m = 2 for biallelic variants and m \u003e 2 for multiallelic variants), we included m − 1 alleles, excluding the most frequent allele as a reference, as independent variables in the regression model. This resulted in the following logistic-regression model:log(odds)=β0+∑j=1m−1β1,jxj+∑k=1K(∑l=1Lβ2,k,lyk,l+β3,kzk)+ε,where β0 is the logistic-regression intercept and β1,j is the additive effect of the dosage of allele j for the variant xj. K and L are numbers of the collections and PCs enrolled in the analysis. yk,l is the lth PC for the kth collection, and zk is the indicator variable for the collection-specific intercept. β2,k,l and β3,k parameters are the effects of yk,l and zk, respectively. An omnibus p value of the variant (pomnibus) was obtained by a log-likelihood ratio test comparing the likelihood of the null model against the likelihood of the fitted model. We assessed the significance of the improvement in fit by calculating the deviance (−2 × the log likelihood ratio), which follows a χ2 distribution with m − 1 degree(s) of freedom.\n\nConditional Association Analysis\nFor conditional association analysis, we considered the regression model including the additional HLA variants as covariates. When conditioning on specific HLA amino acid position(s), we included multiallelic variants of the amino acid residues as covariates. When conditioning on specific HLA gene(s), we included all two- and four-digit classical alleles of the HLA gene(s) (but not alleles with strong correlations [R2 \u003e 0.97]). We consecutively selected the HLA variants to be included as covariates for each HLA gene separately in a forward-type stepwise fashion until no variant satisfied the genome-wide significant threshold (p \u003c 5.0 × 10−8). We tested a multivariate full regression model by including the HLA-C, HLA-B, HLA-A, and HLA-DQA1 risk variants identified by the stepwise regression analysis as covariates and excluding the most frequent allele (or residue) from each locus (or amino acid position) as a reference allele (Table 1). Assuming a PsV prevalence of 2.0%, we estimated phenotypic variance explained by the risk HLA alleles and amino acid polymorphisms on the basis of the effect sizes obtained from the multivariate regression analysis and a liability threshold model.2\n\nTesting for Discordant Effect Sizes on PsA and PsC\nWe tested whether the effect sizes of m classical four-digit alleles of the HLA gene had concordant risks between PsA and PsC, as described elsewhere.26 For each of the two compared phenotypes (PsA-affected versus control individuals and PsC-affected versus control individuals), we calculated multivariate odds ratios (ORs) of m − 1 alleles by including them as binary independent variables in the regression model, where the most frequent allele was excluded as a reference. Let βPsA,1, …, βPsA,m−1 and vPsA,1, …, vPsA,m−1 be the multivariate log ORs and their variances, respectively, in PsA-affected versus control individuals, and let βPsC,1, …, βPsC,m−1 and vPsC,1, …, vPsC,m−1 be those in PsC-affected versus control individuals. We evaluated discordance of the effect sizes between the compared phenotypes (pheterogeneity) by testing the statistic∑i=1m−1(βPsA,i−βPsC,i)2vPsA,i+vPsC,i,which follows a χ2 distribution with m − 1 degrees of freedom under the null hypothesis of concordant effects."}