Fig. 1

a Principal components plot (PC1, PC2) depicting patient diagnosis with number of eosinophils per high-powered field (eos/hpf) in controls (n = 13), and in EoE patients at diagnosis (EoE T0, n = 6) and after treatment (EoE T1, n = 5) after quality control. b Observed variance within all at diagnosis (T0) patient (n = 6) and non-EoE control samples (n = 13) in the first 10 principal components (top panel) against a heatmap showing the correlation between each principal component and phenotype, measured using Kendall’s test statistic for continuous variables and ANOVA for categorical variables (bottom panel). c Clustering of EoE patients at diagnosis (T0) and non-EoE controls (total n = 19) in all CpGs passing quality control using Pearson’s correlation with average clustering. The two principal clusters determined by hierarchical clustering are indicated in green and brown. d Summary of the significant differential methylation analysis results. To be considered as significantly differentially methylated, CpGs needed to have a False Discovery Rate (FDR) p-value < 0.01 and an absolute methylation difference (Δβ) ≥ 0.05. e Heatmap of all samples after quality control, excluding outliers but including biological duplicates (n = 29) subset for the top 25 CpGs significantly differentially methylated between EoE patients at diagnosis (T0) and non-EoE controls (FDR p < 0.01 and Δβ ≥|0.05|) using Pearson’s correlation with average clustering. DMP = Differentially methylated position, DMR = Differentially methylated region. EoE T0 = EoE patients at diagnosis, EoE T1 = EoE patients after first treatment