Fig. 5.

DFNA5 methylation as biomarker for breast cancer. Panel a One CpG in the gene body (CpG12922093) and one CpG in the gene promoter (CpG07504598) were used as predictors. Sensitivity and specificity at various cutoff values for our dataset are shown. The diagonal line represents the line of no discrimination between breast adenocarcinoma and normal breast samples. The predicted probability (cutoff) of 0.87 is indicated with an asterisk. Panel b Three independent datasets, originating from GEO (GSE52865, GSE69914, and GSE60185), were used to validate our model. Two analyses were performed using GSE69914. First, the analysis was performed on 305 breast cancers and 50 normal breast tissues from healthy women (GSE69914_1). Secondly, 305 breast cancers and 42 normal breast tissues, adjacent to the tumor were used to perform the analysis (GSE69914_2). The AUCs for both are almost identical, and the curves are fully overlapping