Computational protocol: Genetic variation in candidate obesity genes ADRB2, ADRB3, GHRL, HSD11B1, IRS1, IRS2, and SHC1 and risk for breast cancer in the Cancer Prevention Study II

Similar protocols

Protocol publication

[…] The selection of specific single nucleotide polymorphisms (SNPs) for this study involved several steps. The list of candidate SNPs was created based on the following criteria: inclusion in the International HapMap Project database []; location within 10 kilobases (kb) of one of the candidate genes (to capture potential regulatory regions) as well as all of the exons and introns; and minor allele frequency of at least 5% in a Caucasian population (to ensure sufficient power). Also included were three nonsynonymous SNPs listed on dbSNP [] that were not in HapMap but had been validated and had a minor allele frequency of above 5% in a Caucasian population (ADRB3 rs4994, IRS1 rs1801276, and IRS2 1805097).These criteria yielded a total of 72 SNPs among the seven genes in this analysis. From this list, tagging were selected using the Tagger program in Haploview (v.3.32) [,]. Tagger is a computer program that is used to select and evaluate tagging SNPs based on the empirical patterns of linkage disequilibrium (LD) called 'bins'. This allows common variation across the region of interest to be captured with fewer SNPs. For this analysis, we used pair-wise tagging to choose SNPs that were correlated at r2 equal to 0.80 or greater with all other SNPs in a LD bin. Furthermore, we required that SNPs previously shown to be associated with cancer and nonsynonymous SNPs be 'forced in' as tagging SNPs. To further reduce the number of SNPs, 27 singleton SNPs (SNPs that were not in LD with any other SNPs) located more than 1 kb from an exon were excluded (singleton SNPs from bins in or near to exons were not excluded). This resulted in a total of 45 SNPs (including the three nonsynonymous SNPs described above) selected. Thirty-nine SNPs were successfully genotyped in the first round. The remaining six failed after two independent attempts, possibly because of high GC content and/or low-complexity in the SNP region.Genotyping was performed by the Center for Medical Genomics at Emory University using the Beckman SNPstream genotyping system. SNPstream is designed to conduct high-throughput, multiplex genotyping using single-base primer extension technology in a tagged fluorescent assay. We used a SNPstream companion primer design website [] to design three primers (two for PCR and one for single-base extension) per SNP and multiplex them into 48-plex primer panels. The PCR primers were used to amplify an approximately 100 base pair region flanking each SNP in a 384-well PCR plate. Image processing and genotype calling were carried out using the GenomeLab SNPstream Genotyping System Software Suite v2.3. Details of the primer sequences and experimental protocol are available upon request.Laboratory personnel were blinded to case-control status. Positive and negative DNA controls were included on each plate and 10% blind duplicates were randomly interspersed among the case-control samples to validate genotyping procedures. Concordance among duplicates samples was above 99%. The overall call rate for each genotyping assay ranged from 87.4% to 99.6%. The two lowest call rates were for IRS2 rs2289046 (87.4%) and IRS2 rs4773092 (92.5%). We carefully evaluated these assays, checked the concordance of duplicate samples, and tested for deviations from Hardy-Weinberg equilibrium. We found no reason to exclude them from the analysis. One SNP (ADRB2 rs1042714) was dropped from the analysis because the allele distribution among controls deviated from Hardy-Weinberg equilibrium (P = 0.004). No other deviations from Hardy-Weinberg equilibrium were observed (at the P < 0.01 level). […]

Pipeline specifications

Software tools Tagger, Haploview
Databases dbSNP International HapMap Project
Application GWAS
Organisms Homo sapiens
Diseases Breast Neoplasms, Neoplasms, Neoplasms, Adipose Tissue