Arya et al [2009] used a logistic regression approach implemented in PLINK to assess the effects of covariates and genotype×sex interactions on the GWAS analysis of RA after accounting for the effects of population stratification. In the single-SNP analysis, only two non-HLA SNPs, on chromosomes 4 and 20, were significant after Bonferroni correction; neither of these SNPs showed a significant genotype×sex interaction. In the genotype×sex interaction analysis, 30 SNP×sex interactions were significant at the uncorrected p<1.0×10−4 level, although none of these survived the stringent Bonferroni correction.
