Computational protocol: Genetic Variation in Iron Metabolism Is Associated with Neuropathic Pain and Pain Severity in HIV-Infected Patients on Antiretroviral Therapy

Similar protocols

Protocol publication

[…] Genomic DNA was isolated from whole blood samples using PUREGENE (Gentra Systems Inc., Minneapolis, MN, USA). Samples were subjected to whole-genome nuclear genotyping using the Affymetrix Genome-Wide Human SNP Array 6.0 platform (Affymetrix, Inc., Santa Clara, CA). Among 579 extracted samples, 560 yielded analyzable genotypes that passed quality-control filters using the Platform for the Analysis, Translation, and Organization of large scale data (PLATO) . Sample genotyping efficiency was 95%. Variants with less than 95% genotyping efficiency or minor allele frequencies (MAFs) less than 1% in the study population were excluded from analysis.Genes reported in the published literature to be associated with iron metabolism and/or neurological phenotypes were searched in the publicly available Online Mendelian Inheritance in Man database (http://www.ncbi.nlm.nih.gov/omim). We then identified 20 genes based on their well-recognized direct or indirect involvement in iron metabolism, transport, storage, or regulation (e.g., the hepcidin pathway). SNPs mapping to the following iron-regulatory and transport genes were selected for this exploratory analysis based on their inclusion in the Affymetrix Genome-Wide Human SNP Array 6.0: HFE, HFE2, SLC40A1, SLC11A1, HAMP, TF, TFRC, TFR2, BMP2, BMP6, CP, SLC11A2, FXN, FTMT, FTH1, ACO1, ACO2, B2M and ATP13A2. Nineteen of these genes were covered by the Affymetrix 6.0 chip, and the platform provided evaluable genotypic information for 192 candidate SNPs (). Many of these iron-related genes/SNPs have also been previously reported to play a role in iron transport within the nervous system, and genes previously linked to neurodegenerative diseases and neural differentiation were included . Based on the common disease-common variant hypothesis, this list was further narrowed to those SNPs with MAFs of at least 5% in one or both of the largest subsets of the CHARTER study population (non-Hispanic blacks or non-Hispanic whites, henceforth referred to as blacks or whites, respectively) . Using a typical analysis strategy that has been applied in numerous previous studies, we used all available genotyped SNPs in those genes, though some of them may be correlated due to LD structure. [...] Statistical tests were performed using publicly available R software (http://www.r-project.org/). For each phenotype, logistic regression was used for single-marker association tests, while adjusting for phenotype-specific covariates in addition to race as a categorical covariate and ancestry PCs, as discussed above , . Additional adjustment for cART-naïve status beyond inclusion of D-drug exposure and HIV viral load did not significantly alter results; hence, this variable was not included in multivariable models. PC variables were computed in advance based on genome-wide genotype data available in all 560 CHARTER study participants. Permutation tests were conducted by randomizing case/control labels in multivariable models while keeping the same numbers of cases and controls as in the original dataset. We generated 1000 permutation datasets. An empirical p-value was computed for each SNP in each phenotype, according to p emp = #{P(π) LocusZoom tool (, generated using HapMap Phase II CEU) .The p-values for all statistical tests were two-tailed, and for this exploratory analysis, the value of α (statistical significance) was set at 0.05 to identify SNPs of potential interest in this exploratory study. A Bonferroni correction for multiple statistical tests was also applied by multiplying p-values obtained for each association test by 192, the number of SNPs evaluated to identify more robust associations. […]

Pipeline specifications

Software tools PLATO, LocusZoom
Application GWAS
Organisms Homo sapiens
Diseases Peripheral Nervous System Diseases, HIV Infections
Chemicals Iron