Computational protocol: Genetic contribution of the leukotriene pathway to coronary artery disease

Similar protocols

Protocol publication

[…] Genomic DNA was extracted from isolated buffy coats using DNeasy isolation kits (Qiagen, Valencia, CA). Genotyping of the 274 haplotype-tagging SNPs was carried out using the Illumina GoldenGate System, which involves using allele-specific primer extension in combination with multiplex PCR with universal primers. Technical details regarding this methodology are available from Illumina, Inc. ( Haplotype-tagging SNPs were selected using the HapMap data for Caucasians and the Tagger program (de Bakker et al. ). Genotyping of previously associated SNPs/variants and those selected for replication in stage 2 was performed using either fragment analysis, as described elsewhere (Dwyer et al. ), or the TaqMan Allelic Discrimination system from Applied Biosystems, Inc. (Foster City, CA) (Livak , ). For determination of HapA and HapK, we genotyped the same SNPs as reported by deCode Genetics (Helgadottir et al. , ), which were rs17222814, rs10507391, rs4769874, and rs9551963 for HapA and rs61937881 (SG12S16), rs2660880, rs6538697, rs1978331, rs17677715, rs2247570, rs2660898, rs2540482, rs2660845, and rs2540475 for HapK (Supplemental Table 1). [...] Prior to analysis, all variants were tested for Hardy–Weinberg equilibrium in subjects without CAD using a χ2 test. SNPs deviating from HWE (p < 0.05) were excluded from further analysis. Haplotypes of ALOX5AP (HapA) and LTA4H (HapK) were estimated using an expectation-maximization (EM) algorithm to generate maximum likelihood estimates of haplotype frequencies, which assigns the probability that each individual possesses a particular haplotype pair. Unconditional multiple logistic regression was used to independently test for association with CAD with adjustment for age and sex under dominant genetic models (as a means to increase sample size and power). The fully adjusted regression models included age, sex, medication use, plasma CRP levels, alcohol consumption, and Framingham ATP-III risk score (which includes smoking and diabetes status) and the results are reported as odds ratios (OR) with 95% confidence intervals (CI). Since we were testing specific hypotheses with respect to the HapA and HapK haplotypes of ALOX5AP and LTA4H, respectively, these analyses were performed as tests of a single haplotype compared to all other haplotypes (i.e., carriers of HapA or HapK vs. non-carriers). For analyses of HapK, we also performed a haplotype score test with all haplotypes having frequencies greater than 1%, as implemented in the Haplo.Stats package. All analyses were performed using SAS version 9.2 (SAS Institute Inc, Cary, NC) or R 2.10.1 ( and carried out separately in Caucasian and African American subjects.Kaplan–Meier failure estimates were plotted to illustrate differences in rates for developing a MACE over 3 years of follow-up as a function of genotype. Subjects experiencing a MACE within 14 days of enrollment were excluded. Due to sample size limitations, we assumed dominant genetic models and significance was assessed using a log-rank test. Relative risk for experiencing a MACE was assessed using Cox proportional hazard models with adjustment for age, sex, medication use, plasma CRP levels, alcohol consumption, and Framingham ATP-III risk score (which includes diabetes status). Other variables tested, but not included in the final model, were BMI and smoking. Adjusted hazard ratio (HR) and 95% CI are reported with 2-sided p values that were considered significant when <0.05. Time-to-event analyses were carried out using Stata 8.2 (StataCorp LP, College Station, TX). […]

Pipeline specifications

Software tools Tagger, haplo.stats
Application GWAS
Organisms Homo sapiens
Diseases Coronary Artery Disease, Myocardial Infarction, Stroke, Atherosclerosis
Chemicals Potassium, Leukotriene A4