Fig. 1

Genome-wide methylation analysis of gastric mucosa and removal of the influence of contaminating leukocytes. a Unsupervised hierarchical clustering analysis of methylation levels of gastric mucosa samples and leukocytes. β values of top 5000 genomic blocks with the highest standard deviation (SD) among all genomic blocks were used. Before correction of leukocyte fractions, never infected young and old gastric mucosae (cluster I) and past-infected and currently infected old gastric mucosae (cluster II) were mixed up (left). After correction (right), cluster II consisted of only past-infected gastric mucosa samples, and the other clusters also had high accordance with the samples. b Fractions of leukocytes in gastric mucosa samples. Fractions of (i) T cells and NK cells (CD4), (ii) B cells (CD19), and (iii) monocytes, neutrophils, and eosinophils (CD14) were separately calculated using specifically methylated CpG sites, and the sum of the three groups was in good accordance with the fraction of pan-leukocytes. A fraction of these cells was used for correction of contaminating leukocytes in gastric mucosa samples