Gene-Gene Associations with the Susceptibility of Kawasaki Disease and Coronary Artery Lesions

Abstract

Kawasaki disease (KD) is a systemic vasculitis primarily affecting children < 5 years old. Genes significantly associated with KD mostly involve cardiovascular, immune, and inflammatory responses. Recent studies have observed stronger associations for KD risk with multiple genes compared to individual genes. Therefore, we investigated whether gene combinations influenced KD susceptibility or coronary artery lesion (CAL) formation. We examined 384 single-nucleotide polymorphisms (SNPs) for 159 immune-related candidate genes in DNA samples from KD patients with CAL (n = 73), KD patients without CAL (n = 153), and cohort controls (n = 575). Individual SNPs were first assessed by univariate analysis (UVA) and multivariate analysis (MVA). We used multifactor dimensionality reduction (MDR) to examine individual SNPs in one-, two-, and three-locus best fit models. UVA identified 53 individual SNPs that were significantly associated with KD risk or CAL formation (p < 0.10), while 35 individual SNPs were significantly associated using MVA (p ≤ 0.05). Significant associations in MDR analysis were only observed for the two-locus models after permutation testing (p ≤ 0.05). In logistic regression, combined possession of PDE2A (rs341058) and CYFIP2 (rs767007) significantly increased KD susceptibility (OR = 3.54; p = 4.14 x 10−7), while combinations of LOC100133214 (rs2517892) and IL2RA (rs3118470) significantly increased the risk of CAL in KD patients (OR = 5.35; p = 7.46 x 10−5). Our results suggest varying gene-gene associations respectively predispose individuals to KD risk or its complications of CAL.

SNP–SNP Interaction Analysis of Soybean Protein Content under Multiple Environments

“…In step 4, each multifactor cell in the n-dimensional space was labeled as high-risk if the values of target traits (e.g., seed length) ≥ average value of target traits. In step 5, we used Pearson χ 2 values to select a model and assess significance (P < 0.001), and the maximum Pearson χ 2 value was used to select the optimization mode (Jiang et al 2009;Kuo et al 2015).…”

“…Selection of the optimization model was based on the χ 2 values. An average Pearson χ 2 value of 10 and a 10-fold cross-validation criterion were used to select the optimization model (Kuo et al 2015). The epistatic interaction effects and their contribution to genetic values and variance were calculated following the method proposed by Cheverud and Routman (1995).…”

“…The role of epistasis in the genetic architecture of quantitative traits has been controversial since early formulations of quantitative genetic theory (Fisher 1930;Wright 1931). Numerous methods to detect epistatic interactions between SNPs have been published, such as exhaustive algorithms (Nelson et al 2001) and MDR (Ritchie et al 2001;Hahn et al 2003;Coffey et al 2004;Moore 2004;Chung et al 2007;Jiang et al 2009;Kuo et al 2015), regression (Heidema et al 2006;Wu et al 2009;Moore et al 2010), heuristic (Carlborg et al 2000;Yang et al 2009), and mutual information (Curk et al 2011) methods, among others (Nelson et al 2001;Purcell et al 2007;Wan et al 2010;Winham et al 2012;Lin et al 2013;Su et al 2015). MDR analysis can effectively reduce genotype predictors from n dimensions to one and incorporates a cross-validation procedure to minimize the rate of false positives.…”

“…de Guia et al (2015) used this technique to reveal the interactions of important gene variants involved in allergies. Many studies using the MDR method have focused on complex human and animal disease-related traits to identify epistatic interactions (Ritchie et al 2001;Hahn et al 2003;Coffey et al 2004;Moore 2004;Chung et al 2007;Jiang et al 2009;de Guia et al 2015;Kuo et al 2015;Rai et al 2015;Li and Sun 2016), but no studies, to our knowledge, have used MDR to analyze soybean quantitative traits.…”

“…In addition, MDR analysis incorporates a cross-validation and (or) permutation procedure to minimize the rate of false positives that may otherwise result from tests involving multiple variables or comparisons (Ritchie et al 2001;Hahn et al 2003;Moore 2004). Although some recent studies on SNP interactions have exploited the MDR method (Ritchie et al 2001;Hahn et al 2003;Coffey et al 2004;Moore 2004;Chung et al 2007;Jiang et al 2009;Kuo et al 2015), these investigations were focused on human and animal complex diseases and related traits. To date, no investigations of SNP interactions involving soybean traits have been conducted.…”