The reference genome is the backbone of all read-mapping analyses in genomics. When sequencing machines produce millions of short DNA fragments, bioinformaticians align those fragments to a reference assembly to determine where each fragment originated. Any sequence absent from the reference cannot be confidently assigned, and reads derived from such regions are typically discarded or misaligned. The most widely used reference, GRCh38, still contains large gaps, particularly in centromeres, pericentromeric satellite arrays, acrocentric chromosome short arms, and other tandemly repeated regions that were technically impossible to resolve with earlier sequencing technologies. The T2T consortium’s complete assembly, designated hs1, closed essentially all of these gaps by combining long-read and ultra-long-read sequencing with advanced assembly algorithms, delivering the first gapless view of a human genome.

The Cologne team reasoned that this gapless reference could change the picture of HPV integration. Viruses are known to preferentially insert into regions of genomic fragility, and highly repetitive DNA is both structurally complex and prone to breakage. If integration events occur inside centromeric or pericentromeric satellites, analyses based on GRCh38 would systematically miss them, not because the events are rare but because the reference lacks the sequence needed to detect them. To test this hypothesis, the researchers reanalyzed datasets from three independent studies that had originally characterized HPV integration using complementary sequencing platforms: short-read Illumina sequencing, PacBio continuous long reads (CLR), PacBio high-fidelity (HiFi) reads, and ultra-long Oxford Nanopore reads. This diversity of technologies allowed them to ask whether the new reference benefits every sequencing modality or only the newest long-read approaches.

The results were striking. Reanalysis against the T2T assembly confirmed the majority of integration breakpoints reported in the original studies, validating the overall reliability of prior work. But in addition to these confirmations, the team identified numerous previously undetected sites, raising the total number of detected breakpoints from 446 to 799. In other words, nearly 40 percent of the integration landscape had been missed by analyses relying on incomplete references. This is not a marginal technical correction; it represents a substantial revision of our understanding of where HPV DNA inserts itself in the cancer genome and how frequently it targets regions that were previously uncharted territory.

Perhaps the most important finding is where the newly discovered breakpoints reside. The additional integrations frequently localized to repeat-rich genomic regions, including centromeric and pericentromeric satellite sequences that were entirely absent from earlier reference assemblies. Centromeres are the structures that anchor chromosome segregation during cell division, and their disruption is a hallmark of chromosomal instability in cancer. Pericentromeric heterochromatin, meanwhile, plays roles in maintaining genome architecture and silencing repetitive elements. Viral integration into these regions could therefore contribute to carcinogenesis through mechanisms that have been almost entirely unexplored, such as compromising centromere function, altering satellite RNA biology, or promoting the formation of dicentric chromosomes and other structural aberrations.

A particularly consequential observation concerns sequencing technology. One might assume that only long-read platforms, with their ability to span repetitive DNA, could detect integrations in satellite regions. The study demonstrates otherwise: integration events within these repeat-rich regions were detectable not only in long-read datasets but also in short-read Illumina data, provided the reads were mapped against the complete T2T reference. The limiting factor was never the sequencing data itself; it was the reference against which the data were aligned. This means that the vast archives of short-read cancer sequencing data already deposited in public repositories contain a wealth of information about viral integration in repetitive DNA that has never been extracted. Reanalysis, rather than new data generation, may be sufficient to uncover clinically relevant events.

The methodological implications extend beyond HPV. Many viruses integrate into host genomes, including hepatitis B virus in hepatocellular carcinoma, HIV in infected T cells, and various oncogenic retroviruses. All integration studies that rely on read mapping share the same dependence on reference completeness. The Cologne findings suggest that any such study conducted against GRCh38 or earlier assemblies may have systematically underestimated integration events, particularly within structurally complex regions. The same logic applies to other rearrangement analyses, such as the detection of extrachromosomal circular DNA (ecDNA), which is increasingly recognized as a driver of oncogene amplification, and to the characterization of common fragile sites (CFSs), which are preferential targets of both viral integration and chromosome breakage.

For HPV research specifically, the expanded integration map opens several avenues of investigation. Previous work established that HPV integrations cluster near or within genes such as TP63, MYC, and E2F-family regulators, and that integration often preserves expression of the viral E6 and E7 oncogenes while disrupting the viral E2 gene that normally restrains them. Whether integrations in centromeric satellites exert oncogenic effects through chromosomal instability, epigenetic perturbation, or other routes remains to be determined. The study also raises questions about formalin-fixed paraffin-embedded (FFPE) clinical specimens, the standard material in pathology archives, whose degraded DNA may behave differently in T2T-based analyses. Tools developed for high-throughput viral integration detection, such as HIVID, will likely need recalibration against the complete reference to realize its full potential.

The authors emphasize that their findings demonstrate a systematic underestimation of HPV integration events caused by incomplete reference assemblies, and they advocate for adopting complete human genome references as the standard in viral integration analyses. Because the study is a brief report based on reanalysis of existing datasets, the biological consequences of the newly mapped integrations remain to be established experimentally, and the researchers are careful not to overstate functional claims. Nevertheless, the message for the field is clear and actionable: the data to find these events already exist, and the reference genome needed to find them is now available. As more laboratories adopt the T2T assembly, the integration landscapes of HPV-associated cancers, and of virus-associated cancers more broadly, are likely to be redrawn with a completeness that was unimaginable only a few years ago, potentially revealing new biomarkers and therapeutic targets hidden in the most repetitive corners of the human genome.

Subject of Research: Reanalysis of HPV genome integration breakpoints using the complete telomere-to-telomere human reference genome

Article Title: Revisiting HPV integration landscapes using telomere-to-telomere human genome reference closes the gaps in HPV integration sites search

Article References: Pacholewska, A., Khan, B. Y., & Schweiger, M. R. (2026). Revisiting HPV integration landscapes using telomere-to-telomere human genome reference closes the gaps in HPV integration sites search. Virology Journal, 23(1), Article 213. https://doi.org/10.1186/s12985-026-03241-y

Image Credits: AI Generated

DOI: 10.1186/s12985-026-03241-y

Keywords: HPV, viral integration, telomere-to-telomere genome, T2T-CHM13, breakpoints, centromeres, satellite DNA, repetitive regions, cervical cancer, genome assembly, bioinformatics, carcinogenesis