Figure 6: Preservation of AID-mediated reprogramming in high-grade serous cancers.

(a, left) 5,000 CpGs with the greatest β increase and decrease in high-grade serous ovarian cancer (n=124) compared with non-serous ovarian cancer types (n=36) (clear cell, endometrioid and mucinous) were selected. (middle) 5,000 CpGs with the greatest median β increase and decrease in fimbrial compartments of BRCA mutation carriers compared with non-carriers were selected (called hyper- and hypomethylated respectively). (right) 5,000 CpGs with the greatest β increase and decrease in primary cultures of FT11 cells force-expressing AID compared with cells transfected with an empty vector were selected. (b) The 5,000 most hypermethylated and 5,000 most hypomethylated CpGs in fimbrial (BRCA mutation carriers versus control) were tested for overlap with the 5,000 CpGs showing the greatest β increase and decrease, respectively, in high-grade serous carcinomas compared with non-serous subtypes (clear cell, endometrioid and mucinous). (c) The 5,000 CpGs showing the greatest β increase and decrease in serous carcinomas compared with other cancer types were tested for overlap with those 5,000 CpGs showing the biggest increase and decrease in methylation, respectively, in cultured cells force-expressing AID compared with cells transfected with empty vector. (d) The 5,000 most hyper- and hypomethylated CpGs in fimbrial samples (BRCA mutation carriers versus control) were similarly tested for overlap with the 5,000 CpGs with greatest β increase and decrease, respectively, in the cultured cells force-expressing AID. All p-values were derived using Fisher’s exact test; all error bars indicate 95% confidence intervals on the estimated odds ratios.