Single-Platform Nanopore Sequencing Enables Diploid Telomere-to-Telomere Genome Assembly and Haplotype-Resolved 3D Chromatin Maps

This study demonstrates that a streamlined, single-platform Oxford Nanopore workflow can generate reference-grade, diploid telomere-to-telomere genome assemblies and haplotype-resolved 3D chromatin maps for diverse individuals, eliminating the need for multi-platform sequencing strategies while achieving high accuracy and scalability.

The Gist

Imagine trying to assemble a massive, 3-billion-piece jigsaw puzzle. For decades, scientists have been trying to put together the complete "instruction manual" for a human being (our genome). The problem? The manual has pages that are ripped, pages that are identical copies of each other, and pages that are written in a code we couldn't read.

Previously, to finish this puzzle, scientists had to use four different types of glue and four different sets of tweezers (different sequencing machines and technologies). It was expensive, slow, and only a few elite labs could afford the tools.

This paper says: "We found a way to do it with just one super-powerful glue gun."

Here is the breakdown of their breakthrough in simple terms:

1. The Old Way: The "Swiss Army Knife" Approach

Before this study, to get a perfect, gap-free copy of a human genome (called a Telomere-to-Telomere or T2T assembly), scientists had to mix and match data from:

• Short-read machines (like a typewriter that prints one letter at a time).
• PacBio machines (a high-precision laser scanner).
• Nanopore machines (a long-read scanner).
• Hi-C machines (a camera that takes photos of how the DNA folds in 3D).

It was like trying to build a house by hiring a bricklayer, a carpenter, an electrician, and a plumber, and then hoping they all agree on the blueprints. It worked, but it was a logistical nightmare.

2. The New Way: The "Master Key" (Nanopore Only)

The researchers at the University of Tübingen discovered they could build a perfect, diploid (two copies of every chromosome, one from mom, one from dad) genome using only Oxford Nanopore technology.

• The Tool: They used a machine that reads DNA strands like a tape recorder reading a very long tape.
• The Trick: They didn't just read short snippets; they used "Ultra-Long" reads. Imagine trying to read a book where the pages are glued together. If you only read a few words at a time, you get lost. But if you can read a whole chapter in one go, you can see exactly where the story repeats and where it changes.
• The Result: They successfully assembled 23 different human genomes using just four "flow cells" (the cartridges that hold the DNA) per person. No other machines were needed.

3. Why "Diploid" Matters: The Two-Track Highway

Most genome maps we have are like a "blended smoothie"—they mix the DNA from mom and dad into one average sequence. But humans have two distinct sets of instructions.

• The Analogy: Think of a highway with two lanes. Sometimes, the left lane has a pothole, but the right lane is fine. If you only look at the "blended" map, you might think the whole road is broken, or you might miss the pothole entirely.
• The Breakthrough: This new method separates the "Mom Lane" from the "Dad Lane" perfectly. They figured out which genetic variants belong to which parent without needing to sequence the parents first. They did this using a special data type called Pore-C, which acts like a "folding map" to see how the DNA loops and touches itself, helping to sort the two lanes apart.

4. The Bonus Features: Methylation and 3D Folding

Because Nanopore technology reads the DNA "raw" (without destroying it), it can see two extra things at the same time:

• The "Highlighter" (Methylation): DNA has chemical tags that act like sticky notes, telling the cell which genes to turn on or off. This method reads those tags directly.
• The "Origami" (3D Structure): DNA isn't just a straight string; it's folded into complex 3D shapes (like a ball of yarn). The researchers mapped how the DNA folds, revealing how different parts of the genome talk to each other.

The Magic: They got the Sequence (the letters), the Phase (Mom vs. Dad), the Tags (on/off switches), and the Folding (3D shape) all from the same four cartridges.

5. What Did They Find?

• Perfect Quality: The quality of their "single-machine" genome is just as good as the "multi-machine" gold standard used by the world's best labs.
• The Hard Parts: They managed to assemble the "impossible" parts of the genome, like the centromeres (the knot in the middle of the chromosome that holds it together) and telomeres (the caps on the ends). These areas are full of repetitive loops that used to be impossible to map.
• Real World Impact: They found that in some people, the "Mom" and "Dad" versions of the genome fold differently, which might explain why some people get certain diseases while others don't.

The Bottom Line

This paper is like saying, "You don't need a team of 10 specialists to fix a car; one mechanic with a really good diagnostic tool can do the whole job."

By proving that one platform can do everything, they have lowered the cost and complexity of creating perfect human genomes. This means:

1. More Data: We can now sequence thousands of people from diverse backgrounds, not just a few.
2. Better Medicine: We can finally see the "hidden" parts of our DNA that cause rare diseases.
3. Democratization: Smaller labs can now do world-class genomics without needing a multi-million dollar budget.

They have turned a "luxury item" (a perfect human genome map) into something that can be produced at scale, opening the door to a new era of understanding human health.

Technical Summary

1. Problem Statement

The completion of the first Telomere-to-Telomere (T2T) human reference genome and subsequent diploid assemblies has revolutionized genomics by resolving previously inaccessible regions like centromeres and segmental duplications. However, current state-of-the-art diploid T2T assembly strategies rely on multi-platform integration, typically combining:

• PacBio HiFi reads for high consensus accuracy.
• Oxford Nanopore (ONT) ultra-long reads for spanning repeats.
• Chromatin conformation capture (Hi-C) or Strand-seq for scaffolding and phasing.
• Parental trio data for phasing (in some cases).

These multi-platform approaches are costly, labor-intensive, and logistically complex, limiting scalability and accessibility to large genome centers. Furthermore, while Nanopore-only strategies have been explored, they often required Duplex sequencing (combining signals from both DNA strands) to achieve high accuracy, which adds complexity and reduces yield. The central question addressed by this study is whether high-quality, reference-grade diploid T2T assemblies can be generated using only standard ONT ultra-long reads and Pore-C data, without PacBio HiFi, Duplex reads, or parental information.

2. Methodology

The authors developed a streamlined, single-platform workflow using Oxford Nanopore Technologies (ONT) on PromethION flow cells.

• Cohort: 23 genetically diverse healthy adults (including two parent-child trios for benchmarking).
• Sequencing Strategy (Per Individual):
  • 3 Flow Cells: Ultra-long (UL) R10.4.1 reads (targeting >100 kb reads) using the Ultra-Long Kit V14.
  • 1 Flow Cell: Pore-C data (chromatin conformation capture) using the Ligation Sequencing Kit V14.
  • Note: No PacBio HiFi, Duplex reads, Strand-seq, or parental sequencing was used for routine assembly.
• Library Preparation:
  • Optimized Ultra-High Molecular Weight (UHMW) DNA extraction to maximize read length.
  • Pore-C libraries prepared via formaldehyde crosslinking, restriction digestion (NlaIII), and proximity ligation.
• Bioinformatics Pipeline:
  • Basecalling: Dorado v1.0.2 (SUP model) with 5mC/5hmC modification calling.
  • Error Correction: Ultra-long reads were error-corrected using HERRO (dorado correct) to generate High-Quality (HQ) reads, serving as a substitute for PacBio HiFi reads.
  • Assembly: Verkko v2.2.1 was used as the primary assembler.
    • HQ reads (corrected UL) were input as --hifi.
    • Raw UL reads were input as --ont.
    • Pore-C data was used for phasing (--porec).
  • Benchmarking: Assemblies were compared against hifiasm (with trio and Pore-C phasing) and multi-platform Human Pangenome Reference Consortium (HPRC) assemblies.
  • Downstream Analysis:
    • Phasing: Pore-C reads were haplotagged to generate chromosome-scale haplotype blocks.
    • 3D Genomics: Haplotype-resolved contact maps were generated to analyze Topologically Associating Domains (TADs).
    • Methylome: Native DNA methylation (5mC) was analyzed simultaneously with sequence.
    • Centromeres: Characterized using CenMap tools.

3. Key Contributions

1. Single-Platform T2T Assembly: Demonstrated that a Nanopore-only workflow (3 UL + 1 Pore-C flow cells) can produce diploid T2T assemblies comparable to hybrid (HiFi + ONT) approaches.
2. Elimination of Duplex Sequencing: Achieved median consensus accuracy of QV50 without using Duplex reads or external polishing, proving that HERRO-corrected R10.4 UL reads are sufficient substitutes for HiFi data in graph-based assembly.
3. Pore-C for Phasing: Validated that Pore-C data alone is sufficient for chromosome-scale haplotype phasing without parental (trio) data, matching the performance of trio-based phasing in terms of switch error rates and scaffold continuity.
4. Integrated Multi-Omics: Successfully reconstructed genome sequence, haplotype phase, DNA methylation, and 3D chromatin topology from a single sequencing run.

4. Key Results

• Assembly Continuity:
  • Across 23 individuals, the study generated 360 gapless chromosomes (34% of all haplotypes) and 446 near-complete T2T scaffolds (42%).
  • The best-performing sample (T2T17) achieved 31 gapless T2T contigs and 8 near-complete scaffolds.
  • Autosomal chromosomes 1–12 and 16–20 were consistently resolved to T2T contig/scaffold levels. Acrocentric chromosomes (13, 14, 15, 21, 22) remained challenging but were resolved in specific individuals.
• Accuracy Metrics:
  • Consensus Accuracy: Median QV of 51 (approx. 6 errors per megabase) across k-mers (21, 31, 37), comparable to HPRC hybrid assemblies.
  • Structural Integrity: Low rates of flagged mis-assemblies (verified by NucFlag and Flagger), with errors enriched only in known difficult loci.
  • Gene Completeness: Distributions of missing multi-copy genes (MMC) and single-copy genes (MSC) were comparable to HPRC assemblies.
• Phasing Performance:
  • Pore-C-based phasing produced chromosome-scale blocks with switch error rates and scaffold continuity comparable to trio-based phasing.
  • Assemblies using hifiasm showed higher switch errors and lower continuity compared to Verkko, despite slightly higher contig counts.
• Functional Genomics:
  • 3D Chromatin: Generated haplotype-resolved contact maps revealing allele-specific TAD boundaries at the imprinted IGF2-H19 locus and clear signatures of X-chromosome inactivation in female samples.
  • Centromeres: Resolved 2–42 centromeres per individual, revealing extensive variation in High-Order Repeat (HOR) organization and length.
  • Methylation: Simultaneously mapped allele-specific methylation patterns, including imprinting and X-inactivation.

5. Significance

• Scalability and Accessibility: By removing the need for PacBio HiFi, Duplex sequencing, and parental data, this workflow significantly reduces cost, laboratory complexity, and hands-on time. This makes T2T genomics accessible to smaller labs and enables population-scale studies previously limited by resource constraints.
• Reference-Grade Quality: The study proves that Nanopore-only assemblies can achieve reference-grade quality (QV50, high continuity) suitable for the Human Pangenome Reference Consortium (HPRC).
• Integrated Functional Insights: The ability to derive sequence, methylation, and 3D structure from a single platform provides a unique, unified view of the genome. This is crucial for understanding how epigenetic and structural variations (e.g., in TADs) contribute to disease, particularly in regions inaccessible to short-read sequencing.
• Future Impact: The authors plan to expand this resource to 50 publicly available T2T genomes, providing a foundational dataset for studying structural variation, centromere evolution, and regulatory architecture across diverse populations.

In conclusion, this paper establishes a new standard for diploid genome assembly, demonstrating that a single sequencing platform (ONT) is sufficient to generate comprehensive, phased, and functionally annotated T2T genomes, thereby democratizing access to the most complete view of human genetic diversity.