Statistical analysis
For data filtering and finding significant SNPs, we used PLINK version 1.07, that is a tool made for analyzing whole-genome association using computational methods [11]. We used the default options of PLINK [11], and we analyzed phenotypes by logistic regression test for classifying patients and normal subjects and estimating factors. We also assessed the result factors of the logistic regression by Student's t-test for revealing meaningful differences between patients and normal subjects using R version 3.0.2 for finding gastritis-associated factors. Then, we used the receiver operating characteristic (ROC) curve and area under the curve (AUC) scores to confirm the prediction ability of the factors.