Statistical analysis
The distribution of the included variables of our study was tested via Shapiro- Wilk method. Normal distribution data was presented as mean ± standard variance, otherwise median and (interquartile range) was presented. Correspondingly, student’s t test or Mann-Whitney U test were used to test the between group difference of the mean or mean rank, respectively. Spearman’s rank correlation efficiency was test to evaluate the association between ERAP1 methylation level and other continuous or hierarchical variables. Univariate logistic regression analysis was used to calculate the odds ratio (OR). Odds ratio (OR) with 95% confidence interval (CI) and forest plot were used to evaluated association of individual methylation site and AS visualized. Multiple logistic regression analysis was used to establish a regression model. Receiver operating characteristic (ROC) curve and area under curve (AUC) were applied to assess the predictive value of the ERAP1 methylation as a biomarker of AS. All these data analysis diagram plotting were accomplished by SPSS 23.0 software (SPSS Inc., Chicago, IL, USA), GraphPad Prism 7.00 (GraphPad Software Inc., CA, USA) and R software. A two-tailed P < 0.05 was taken as statistically significant, and multiple comparisons were adjusted through Bonferroni method.