48 Chapter 2 root values of the ΔCq, and p values were two-sided, with 0.05 as significance threshold. Statistical analyses were performed with R Statistical Software (v3.5.1) in RStudio, using the following R-packages: pROC (v1.13.0), gplots (v3.0.1), ggplot2 (v3.1.0) and glmnet (v2.0-16). RESULTS METHYLATION LEVELS AND CIN3 DETECTION BY SINGLE METHYLATION MARKERS Methylation levels of all six methylation markers increased significantly with increasing underlying disease severity, from controls to CIN3 and cervical cancer (Kruskal-Wallis test with post hoc pairwise Wilcoxon-Mann-Whitney U-tests, all p values < 0.001; Figure 2.1). A significant positive association between all six markers was observed within the group 0 5 10 15 20 Control CIN3 Cancer Methylation level (sqrt transformed ΔCq) ASCL1 LHX8 GHSR SST ZIC1 ST6GALNAC5 Figure 2.1 Methylation levels increase with severity of cervical disease. DNA methylation levels of six single methylation markers relative to that of the housekeeping gene ACTB (square-root transformed ΔCq; Y-axis) in HPV-positive cervical samples corresponding to different histological categories of underlying cervical disease are shown. Boxplots are based on a five-number summary (i.e., minimum, first quartile [Q1/25th percentile], median [Q2/ 50th percentile], third quartile [Q3/75th percentile] and maximum with outliers displayed as single points). Differences between histological categories were significant: Kruskal-Wallis test p value < 0.001, pairwise Wilcoxon-Mann-Whitney U-test p values < 0.001. Abbreviations: CIN, cervical intraepithelial neoplasia; Cq, quantification cycle; HPV, human papillomavirus; sqrt, square-root

RkJQdWJsaXNoZXIy MjY0ODMw