High-grade serous ovarian carcinoma is one of the most genomically chaotic cancers known to medicine, and yet the picture scientists hold of its genome has been built with instruments that cannot see whole stretches of it. A new mini-review published in Frontiers in Molecular Biosciences argues that the missing pieces—centromeres, telomeres, transposable elements and extrachromosomal DNA—are not marginal curiosities but may form an additional layer of genome architecture that conventional sequencing has systematically overlooked. The authors, Xu, Ren and Yuan of Zibo 148 Hospital and Qilu Hospital of Shandong University, synthesize emerging long-read evidence and caution that, so far, that evidence rests on remarkably small numbers of patients.
The established framework for this disease comes from large-scale efforts such as The Cancer Genome Atlas, which analyzed 489 high-grade serous ovarian carcinomas and found TP53 mutations in 96 percent of cases, homologous recombination defects in roughly half, and widespread copy-number alterations. Later whole-genome studies added structural detail: gene-disrupting rearrangements affecting RB1, NF1, RAD51B and PTEN, amplification of CCNE1, and BRCA1/2 reversion events have all been linked to chemoresistant disease. Copy-number signature analyses of 117 tumors, validated in 527 more, uncovered seven reproducible patterns reflecting distinct mutational processes, while rearrangement signature work identified five recurring structural patterns, one of which was associated with poorer overall survival.
Deep whole-genome sequencing of 324 tumors has since sharpened this evolutionary picture. That study identified frequent homologous recombination deficiency and whole-genome duplication, alongside chromothripsis, breakage-fusion-bridge events, chromoplexy and extrachromosomal DNA. Two broad trajectories emerged: one dominated by homologous recombination deficiency and deletion-rich genomes, the other by whole-genome duplication with an enrichment of duplications, chromothripsis, breakage-fusion-bridge events and ecDNA. Approximately 16 percent of tumors in that cohort harbored predicted ecDNA, a subset carrying amplified oncogenes, and at the CCNE1 locus diverse processes converged on high-level amplification. The lesson, the reviewers stress, is that extreme structural instability in this cancer cannot be attributed to homologous recombination deficiency alone.
What all of these studies share is a technical blind spot. Exome sequencing, microarrays and short-read whole-genome sequencing detect point mutations and copy-number changes well, but they map repetitive sequences poorly and can fragment complex rearrangements into disconnected breakpoint calls. Centromeres, telomeres, segmental duplications and circular DNA remain underrepresented. Long-read sequencing can span repetitive regions and rearrangement junctions, and nanopore platforms additionally read DNA methylation directly on individual molecules, enabling haplotype-resolved epigenomic profiling. The telomere-to-telomere human reference assembly, T2T-CHM13, adds nearly 200 megabases of sequence absent from earlier references, including complete centromeric satellite arrays.
The first of the two pivotal long-read studies applied Oxford Nanopore sequencing with the T2T-CHM13 reference to six pretreatment high-grade serous ovarian carcinomas and matched blood samples. Every tumor showed pronounced hypomethylation across centromeric satellite regions compared with matched blood. Centromeric methylation profiles also distinguished tumors with high versus low homologous recombination deficiency scores, with more substantial hypomethylation in the low-HRD group, and reduced or variable alpha-satellite read coverage pointed to structural heterogeneity within centromeric arrays. The same study found pronounced telomere shortening at the level of individual chromosome arms and allele-specific methylation of the TERT hypermethylated oncological region in three of six tumors—an epigenetic pattern that bulk methylation assays would miss. The authors of the review emphasize that the small cohort and the lack of tissue-matched normal fallopian tube controls mean these results do not establish a causal link between centromeric hypomethylation and chromosomal instability; they suggest a facet of the instability phenotype whose interplay with homologous recombination deficiency may be more nuanced than assumed.
Transposable elements form the second pillar of the hidden genome. LINE-1 is the only autonomous, protein-coding retrotransposon family still active in humans, and its derepression appears early in serous carcinogenesis. In one immunohistochemistry study spanning 12 p53 signatures, 32 serous tubal intraepithelial carcinomas and 112 ovarian carcinomas, none of the p53 signatures showed substantial LINE-1 ORF1p expression, whereas strong expression was frequent in precursor lesions and correlated with LINE-1 promoter hypomethylation, persisting in invasive carcinomas. A follow-up study detected ORF1p in tumors and tubal precursors but not in benign fallopian tube epithelium, and found the protein released into conditioned medium, ascites and plasma, making it a candidate early-event biomarker—though the cross-sectional design of the earlier work precludes firm conclusions about timing. The nanopore study added an epigenomic dimension, reporting pronounced hypomethylation of L1HS and endogenous retroviral elements, with more modest changes at AluY and SVA elements.
Retrotransposition is not merely a molecular fossil record; it is ongoing. A recent analysis of de novo LINE-1 insertions in clinical samples, using whole-genome sequencing, long-distance inverse PCR and nanopore sequencing across two cohorts of 50 tumors from 28 patients and 57 tumors from 38 patients, found considerable inter-patient variability in insertion burden. Among patients with high LINE-1 burden, within-patient tumor pairs shared fewer insertion sites than single-nucleotide variants, as quantified by the Jaccard index—a comparison the reviewers note is descriptive rather than a measure of equivalent mutation rates. The site-specific insertions nonetheless provide direct evidence of continuing retrotransposition and intrapatient heterogeneity in advanced disease. High-burden tumors showed proliferative characteristics, while those with little detectable insertion activity were comparatively enriched for immune response and cell death pathways.
Transposable element activation also has consequences beyond insertion. Transcriptomic and chromatin profiling in ovarian cancer cell-line models showed that cisplatin exposure and resistance alter the expression of LINEs, Alu elements and endogenous retroviruses, with TE transcription often correlating with nearby host genes and producing chimeric TE-gene transcripts. Separately, work in fallopian tube premalignant lesions and experimental ovarian cancer models demonstrated that loss of p53 permits transcription of immunogenic repetitive elements, inducing chronic viral-mimicry signaling. Rather than eliminating the tumor, this chronic exposure appeared to select for cells with enhanced tolerance to cytosolic nucleic acids and reduced immunogenicity—a proposed route to early immune evasion that remains a model rather than a proven mechanism in patients.
Extrachromosomal DNA ties several of these threads together. An early Circle-seq study in four paired primary and metastatic tumors found abundant eccDNA with distinct profiles at each disease site, but short reads cannot reconstruct large circular amplicons. Long-read whole-genome sequencing of primary and paired metastatic tumors from 12 patients, alongside six normal fallopian tube samples, showed that ecDNA is more genomically unstable and less methylated than chromosomal DNA. These ecDNAs carried immunomodulatory genes and extensive enhancer sequences; circular rearrangements enabled enhancer hijacking and mobile-enhancer mechanisms, and transposable elements were activated on hypomethylated ecDNA, with that activity correlating with epithelial-mesenchymal transition-related gene expression and adverse clinical outcomes. Whether ecDNA directly promotes transposable element activation or both stem from a shared hypomethylated state remains unresolved.
The reviewers are candid about limits. Genome-wide long-read evidence in this cancer currently derives primarily from the six-patient and 12-patient studies, and clinical applicability of any of the proposed biomarkers—ORF1p assays, tumor-specific LINE-1 insertions, ecDNA-derived sequences, centromeric or telomeric methylation patterns—remains unproven. Long-read workflows demand high-molecular-weight DNA, may miss low-frequency subclonal events at typical depths, and still rely on immature callers for complex structural variants and ecDNA reconstruction, with even T2T references vulnerable to mapping ambiguity in the most repetitive regions. Therapeutic ideas, such as exploiting DNA damage response dependencies in tumors with high LINE-1 activity or targeting ecDNA maintenance, are explicitly framed as mechanistic hypotheses. The proposed priorities are larger cohorts, tissue-matched controls and sampling across the disease’s developmental arc from fallopian tube epithelium through precursor lesions to invasive carcinoma, to determine whether the hidden genome shapes diagnosis, risk stratification or treatment.
Subject of Research: Long-read sequencing evidence for repetitive DNA, extrachromosomal DNA and complex structural variation in high-grade serous ovarian carcinoma
Article Title: The hidden genome of high-grade serous ovarian carcinoma: repetitive DNA, extrachromosomal DNA, and complex structural variation
Article References: Xu, X., Ren, X., & Yuan, Z. (2026). The hidden genome of high-grade serous ovarian carcinoma: repetitive DNA, extrachromosomal DNA, and complex structural variation. Frontiers in Molecular Biosciences, 13, 1988095. https://doi.org/10.3389/fmolb.2026.1988095
Image Credits: AI Generated
DOI: 10.3389/fmolb.2026.1988095
Keywords: high-grade serous ovarian carcinoma, long-read sequencing, extrachromosomal DNA, transposable elements, LINE-1, centromeres, telomeres, structural variation, DNA methylation, genome instability, ecDNA, T2T reference genome
News Source: Nathaniel Bowman. (October 28, 2026). Long-read sequencing reveals the hidden genome of ovarian cancer. Scienmag.



