The squared correlation coefficient (r) between SNPs was used as a measure of LD.
Whereas some of these imputation methods use linkage analysis to exploit known relationships between individuals, in many cases, knowledge of relationships is not required and population-wide LD between SNPs is used.
For genotype interaction analysis, all possible genotype combinations between selected SNPs were used as an independent variable, using binary logistic regression with the forward likelihood ratio method, keeping the gender as a covariate.
The block-wise P-values are then combined by the same extended Simes test again to produce the multivariate gene-based P-value PMGAS in which the correlations between the key SNPs are used to correct for the dependency of the K block-wise P-values.
As an initial step, all pairwise epistatic interactions between SNPs were evaluated using ' information gain' – a metric used in Information Theory.
Distributions of demographic and clinical variables were compared between patients and controls using chi-square tests or t-tests, and estimates of pair-wise LD between SNPs were obtained using Haploview software, version 4.1 [30].
The 13 SNPs shown in Figure 1 were tested for Hardy-Weinberg equilibrium using Haploview 3.2 software [ 13] and estimates of linkage disequilibrium (LD) between SNPs were calculated using D' and r[ 17], to select one SNP that represents each LD block or an unlinked area.
Interaction effects between SNPs were modeled using four mutually exclusive levels: (1) subjects without either risk allele, (2) subjects with one risk allele in the first SNP but without the risk allele in the second SNP, (3) subjects without the risk allele in the first SNP but with the risk allele in the second SNP, and (4) subjects with both risk alleles.
Estimates of linkage disequilibrium (LD) between SNPs were calculated using D' and r.
To identify relationships between the SNPs that were used in the study and reduce the burden of tests because some tests were not independent, a linkage disequilibrium (LD) analysis was performed for 300 samples using Haploview v. 4.1 [ 27].
