Humanhap550 beadchip

DNA extraction and genotyping were performed as described before,^{9 (link)} on saliva samples by National Genetics Institute (NGI, Los Angeles, CA, USA). Samples have been genotyped on one of four genotyping platforms. The V1 and V2 platforms were variants of the Illumina HumanHap550+ BeadChip (Illumina, San Diego, CA, USA), including about 25 000 custom single-nucleotide polymorphisms (SNPs) selected by 23andMe, with a total of about 560 000 SNPs. The V3 platform was based on the Illumina OmniExpress+ BeadChip, with custom content to improve the overlap with the V2 array, with a total of about 950 000 SNPs. The V4 platform in current use is a fully custom array, including a lower redundancy subset of V2 and V3 SNPs with additional coverage of lower-frequency-coding variation, and about 570 000 SNPs. Samples that failed to reach 98.5% call rate were reanalyzed. Participant genotype data were imputed against the September 2013 release of 1000 Genomes^{15 (link)} Phase1 reference haplotypes, phased with ShapeIt2. Prior to imputation, we excluded SNPs with Hardy–Weinberg equilibrium P<10⁻²⁰, call rate <95%, or with large allele frequency discrepancies compared to European 1000 Genomes reference data. Additional details on the imputation procedure could be found in Supplementary Text S1.

DNA extraction and genotyping were performed, as described previosuly^{6 (link),13 (link)}. Briefly, samples were genotyped on platform variants (V1 and V2) of the Illumina HumanHap550 + BeadChip (Illumina Inc., San Diego, CA), and included ~ 25,000 custom single nucleotide polymorphisms (SNPs) selected by 23andMe, with a total of ~ 560,000 SNPs. A custom content platform (V3) based on the Illumina OmniExpress + BeadChip was used to improve the overlap, with a total of ~ 950,000 SNPs. A fully custom array platform (V4) was used which included a subset of SNPs with additional coverage of lower-frequency coding variation, and ~ 570,000 SNPs. The samples that failed to reach 98.5% call rate were re-analyzed. Prior to imputation of genotype data against the September 2013 release of 1000 Genomes^{14 (link)} Phase1 reference haplotypes, we excluded SNPs with Hardy–Weinberg equilibrium p < 10⁻²⁰, call rate < 95%, or with large allele frequency discrepancies compared to European 1000 Genomes reference data^{15 (link)}. Additional details on the imputation procedure are provided in Supplementary Text S1.

The GAIN and TGEN samples were genome-wide genotyped on the Affymetrix 6.0 SNP array. For the BoMa sample, the Illumina HumanHap550 BeadChip was used. All genotypes were imputed based on 2.1 million HapMap Phase 2 markers (McMahon et al. 2010 (link)). Due to computational runtime constraints, our analysis is based on a selected number of markers. We included only those SNPs that showed an association p-value of less than 0.001 in a recent meta-analysis of 4961 BD patients and 7294 controls (Additional file 4: Text S1, Methods-SNP selection). Our resulting SNP set comprised 5487 SNPs, on which LD pruning (Additional file 4: Text S1, Methods-Linkage disequilibrium) was performed in order to reduce redundancy within the genotype data before the discovery step and to decrease runtime. This left us with a total of 1599 SNPs. Of these, 1581 SNPs were available in all three samples studied. As the ARM approach requires binary variables we had to transform the genotype information into a binary format (Additional file 4: Text S1, Methods-Genetic Models).

Study participants included 121 COPD cases genotyped as part of the ECLIPSE study [36 ] using the genome-wide Illumina HumanHap550 BeadChip. Each participant had data on ~6.1 million SNPs, either directly genotyped or imputed using the 1000 Genomes EUR reference panel (March 2010) [48 (link)]. Details on quality control assessment, filtering of the SNPs and genotype imputation have been described elsewhere [49 (link), 50 (link)]. Gene expression was measured from whole blood samples on the Affymetrix HGU 133Plus 2.0 chip, and eQTL analysis was performed as previously described [8 (link)]. Gene expression data were log-transformed and quantile normalized using the RMA function in the “affy” R package [51 (link)] (step1 in Fig 1). We then regressed expression values on age, gender, the first principal components derived from the genotype data on all ECLIPSE participants [52 (link)], and the first 13 principal components from gene expression, retaining residuals from this regression for further eQTL analysis (step2 in Fig 1). Genotype data of all samples used in this study are available in dbGap (phs001252.v1.p1, in process). ECLIPSE expression data has been previously submitted to GEO as part of another project (229 samples) with GSE76705. Note that the ECLIPSE dbGaP submission (phs001252.v1.p1, in process) will also contain links to the GEO expression data.