Endometriosis affects 10% of women of reproductive age and is associated with an increased risk of developing specific ovarian cancer histotypes, such as clear cell ovarian carcinoma (CCOC) and endometrioid ovarian carcinoma (EOC). While most research has focused on exploring direct genomic associations involving cancer-driver genes with inconclusive results, recent genome-wide association studies have delved into non-coding regions to identify additional genetic factors. These studies, however, face challenges in detecting rare and somatic mutations. Advances in next-generation sequencing, accessible datasets, and deep-learning models have enabled comprehensive genomic studies. This study aimed to identify specific non-coding mutations in ovarian endometriosis (OE) and ovarian cancer (OC) that could influence disease development. Using whole exome sequencing (WES) data from 20 OE and OC specimens and a variant calling pipeline called Sarek, 11774 somatic non-coding mutations were identified. Of these, 49.2% (5794 variants) were predicted to be within regulatory elements. Furthermore, 50 common mutations were found in more than three specimens. Unsupervised clustering grouped the results into three categories: the OE group (9 samples) which had specific variants in the LRRC4B, REPS1, and SLC1A1 genes, while variants in CNN2P1, ANKRD20A4P, and ZNF806 were found in both OE and OC groups (20 samples). Additionally, 44 mutations were found in 3 or more OC samples (out of 11), with VGLL1 and PRSS56 being the most recurrent across both OC histotypes. In summary, this study has identified specific common mutations prioritized based on their frequency and functional annotation. However, understanding their functional significance will need further experimental validation.