Back

Spatio-Temporal Landscape of Whole-Genome DNA Methylation Patterns in Ovarian Cancer

Marchi, G.; Lavikka, K.; Li, Y.; Isoviita, V.-M.; Micoli, G.; Afenteva, D.; Pöllänen, E.; Holmström, S.; Häkkinen, A.; Valkonen, E.; Maarala, I.; Astren, J.; Facciotto, C.; Dietlein, F.; Hietanen, S.; Muranen, T.; Oikkonen, J.; Hynninen, J.; Virtanen, A.; Lahtinen, A.; Hautaniemi, S.

2026-03-25 genetic and genomic medicine
10.64898/2026.03.22.26348877 medRxiv
Show abstract

Understanding how epigenetic mechanisms shape metastatic dissemination and therapy response in ovarian high-grade serous carcinoma (HGSC) has been hindered by the absence of large, anatomically diverse, and longitudinal datasets. Here, we present a comprehensive resource of the HGSC methylome, comprising 404 whole-genome bisulfite sequenced samples from 125 patients belonging to the real-world DECIDER trial. Tumor-specific methylomes were integrated with matched RNA sequencing, whole-genome sequencing and clinical outcomes, which enabled identification of > 8,500 regulatory promoters. Our results show that the regulatory methylome is largely established early in tumor progression and remains stable through chemotherapy. Tumors with poor response to treatment are marked by widespread promoter hypermethylation, particularly cancer cells in ascites, leading to epigenetic silencing of pathways targeted by chemotherapy agents. Our proof-of-principle analysis shows that these epigenetic signatures are detected in plasma-derived DNA, highlighting the value of DNA methylation in liquid biopsy-based disease monitoring. This study provides a foundational resource for dissecting the spatial and temporal dynamics of cancer methylomes and establishes a framework for integrating epigenetic states with clinical trajectories in HGSC.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.