Expanding reference genomes with T2T assembly to reflect global diversity

High-quality reference genomes are the backbone of precision medicine. However, despite recent efforts by the scientific community to generate more diverse genetic data, the reference genomes available still lack representation across global populations. Furthermore, due to the limitations of legacy sequencing approaches, many assemblies remain incomplete. Now, researchers are demonstrating how telomere-to-telomere (T2T) assemblies, generated exclusively with Oxford Nanopore sequencing, could help close both of these gaps.

A local genome with global significance

At London Calling 2025, Prasad Sarashetti (A*STAR Genome Institute of Singapore, Singapore) shared results from a project aimed at assembling complete, high-quality T2T genomes using only Oxford Nanopore sequencing data without the need for a hybrid approach via multiple technologies. Prasad highlighted the recent advancements made possible by long reads, noting their essential role in producing the first complete T2T genome assembly in 2022¹.

Prasad announced the completion of the team’s first Indian genome — an important milestone in addressing the gaps in the availability of diverse reference genomes for precision genomics². This genome assembly was constructed using a hybrid approach, combining PacBio HiFi reads with ultra-long Oxford Nanopore reads, and achieved a quality value (QV) of approximately 82 — equating to one error in every 100 million bases, making it one of the highest quality diploid genomes available to date. Next, the team wanted to push the boundaries further by asking: could they achieve the same high standards with Oxford Nanopore data alone?

The benefits of an Oxford Nanopore-only approach

To answer that question, the team assembled three human genomes, using both a hybrid method — PacBio HiFi reads combined with Oxford Nanopore ultra-long reads — and an Oxford Nanopore-only pipeline, which utilised Pore-C chromatin conformation capture for phasing.

Although contiguity values (NG50s) were similar across most of the assemblies, the Oxford Nanopore-only assemblies consistently outperformed those assembled using the hybrid approach in terms of the number of complete chromosome-length sequences (Figure 1). The exception in this instance was the acrocentric chromosomes, which remain notoriously challenging due to the presence of ribosomal DNA (rDNA) clusters.

Figure 1. Assemblies generated from Oxford Nanopore-only data featured higher numbers of T2T chromosomes than those generated using a hybrid approach. Figure from Prasad Sarashetti, reproduced with permission.

The de novo assemblies were also evaluated for structural accuracy and completeness. The team used a technique which identified how many kmers were present in the assembly, but not in the original reads, as an indication of consensus errors. Using this method of testing assembly quality, the Oxford Nanopore-only assemblies showed comparable or better completeness than the hybrid methods and exhibited fewer structural errors, suggesting a highly faithful representation of the original genome.

Enabling population-specific insights

The utility of Oxford Nanopore reads of unrestricted length, and specifically ultra-long reads, became even more apparent when the team evaluated variants in the Indian genome assembly. Prasad highlighted one particularly impactful finding: a structural variant (SV) in a repetitive region that was absent from African and European reference genomes but appeared to be a fixed allele in South Asian populations.

Referring to a curated list of genes with known clinical relevance, Prasad noted that ‘this SV falls within the ACMG gene list’ and that ‘proper or accurate reconstruction of this region would enable us to see this SV, which is … [a] fixed allele in South Asian genomes’. This discovery not only underscored the power of ultra-long Oxford Nanopore reads to resolve challenging genomic regions, but also pointed to the clinical importance of population-specific references.

Scaling with confidence

For researchers looking to scale up high-quality genome sequencing, computational cost is a key consideration. The team evaluated an emerging CPU-based method recently introduced in the hifiasm^3–5 assembler and found that it performed as well as existing, more resource-intensive GPU-based tools, showing promise for a more efficient computational workflow.

Reflecting on the team’s research, Prasad concluded: ‘yes, we can generate reference-quality assemblies only using [Oxford Nanopore] platforms … [and] given the improvement we have seen, it is worth to explore this path for large-scale projects’.

A path forward for diverse reference genome assembly

The team’s findings provide compelling evidence that Oxford Nanopore sequencing can stand alone in delivering complete, reference-quality, phased human genome assemblies. For scientists aiming to characterise genomes at scale — and to close longstanding diversity gaps in human genomics — these results offer a practical and powerful path forward.

As the global scientific community continues to prioritise population-specific reference genomes, studies like this one illustrate how this accessible, scalable, and accurate technology can unlock a more inclusive era of precision medicine.

Oxford Nanopore Technologies products are not intended for use for health assessment or to diagnose, treat, mitigate, cure, or prevent any disease or condition.