The work was carried out by the Telomere-to-Telomere, or T2T, Consortium using HG002, a human genome sample obtained from a living donor and widely used as a reference material by sequencing and diagnostic laboratories. The researchers assembled each chromosome from one telomere, the protective structure at one end, to the other, producing two separate chromosome sets that represent the maternal and paternal genomes. This is technically more difficult than assembling a single genome because the two copies are highly similar but not identical. Computational systems must determine which DNA fragments belong to which parental chromosome while preserving every small difference between them.

The new sequence adds more than 900 million DNA letters that were absent from previous benchmarks and reveals roughly 15% more of the genome than the earlier reference standard. These newly accessible regions include parts of both sex chromosomes, highly repetitive stretches, and sequences containing genes and regulatory elements that may influence disease risk. Such regions have often been excluded from clinical sequencing because standard technologies struggled to read them accurately or because researchers could not determine their correct position within the genome. By resolving these difficult segments, the T2T approach could expose genetic variants that have remained invisible in routine testing.

The advance builds on the consortium’s landmark 2022 completion of the first truly complete human genome sequence. That project filled in approximately the final 8% of a single reference genome, including many repetitive regions and the previously incomplete Y chromosome. The new effort goes further by applying improved sequencing platforms, assembly algorithms and validation methods to a diploid genome. Long-read sequencing technologies were central to the work because they generate DNA fragments thousands or even millions of letters long, allowing researchers to span repetitive sequences that would be broken into ambiguous pieces by older short-read methods.

Accurately assigning genes to each chromosome copy was another essential part of the project. Scientists at Johns Hopkins led by computational biologist Steven Salzberg analyzed the two chromosome sets to identify and annotate their genes, while Michael Schatz’s laboratory contributed to extensive validation of the assembly. Independent checks were used to test whether the reconstructed sequence contained errors, missing segments or incorrectly joined fragments. The result is intended not only as a biological reference but also as a measurement standard for companies developing DNA sequencing instruments, analysis software and clinical diagnostics.

Researchers say the development could change the logic of medical genomics. Current clinical analyses generally search for variants that differ from a standard reference genome. This strategy can perform well when a patient’s DNA resembles the reference, but it becomes less reliable in genomic regions where the reference is incomplete or structurally different. A complete genome assembled for each patient would instead provide an individualized baseline. Genetic analysis could then examine substitutions, insertions, deletions, duplications and larger rearrangements across the entire sequence without automatically discarding regions that do not align well with the traditional reference.

The immediate medical benefit could be improved diagnosis for children and adults with rare genetic disorders. Genome sequencing is already used in such cases, but more than half of patients may still leave testing without a clear molecular explanation. Missing or misread regions can conceal the mutation responsible for disease, particularly when it lies in a repetitive sequence or involves a complex structural change. A complete diploid assembly could help clinicians identify these causes more accurately, potentially ending years of uncertainty for families and guiding treatment, monitoring and reproductive decisions.

The same approach may eventually strengthen predictions for common diseases. Variants in the BRCA1 and BRCA2 genes are already used to estimate breast cancer risk, but researchers believe that many additional risk-associated changes remain undiscovered in difficult-to-sequence portions of the genome. More complete reference data could improve studies of cancer, cardiovascular disease, immune disorders and neuropsychiatric conditions. When combined with genomes from large and diverse populations, these sequences could also support artificial intelligence models trained to recognize disease-related patterns while reducing the bias created by relying on a single, historically limited reference genome.

The consortium estimates that a complete and highly accurate human genome can now be generated for about $5,000, compared with the roughly $5 billion, in current dollars, spent on the Human Genome Project, which concluded in 2003. Although routine whole-genome sequencing still raises questions about privacy, data storage, consent and the interpretation of uncertain findings, the technical barrier is rapidly falling. The researchers envision a future in which a person’s complete genome is sequenced early in life, securely linked to medical records and revisited as scientific knowledge improves. The study is part of a broader package of work in Cell and Cell Genomics that also presents complete or near-complete genomes for macaques, marmosets, zebra finches, rats, voles, horses, donkeys and giraffes, extending the same genomic precision to research on evolution, biodiversity, agriculture and animal health.

Subject of Research: Complete diploid human genome sequencing and personalized genomics

Article Title: Complete, high-quality diploid human genome reconstructed from telomere to telomere

Web References:
https://engineering.jhu.edu/faculty/adam-phillippy/
https://hub.jhu.edu/2022/03/31/johns-hopkins-scientists-first-complete-sequence-human-genome/
https://www.bme.jhu.edu/people/faculty/steven-l-salzberg/
https://engineering.jhu.edu/faculty/michael-schatz/

References:
Cell, DOI: 10.1016/j.cell.2026.06.016

Keywords

Human genome sequencing, diploid genome, Telomere-to-Telomere Consortium, personalized genomics, genetic disease diagnosis, long-read sequencing, genomic medicine, structural variants, precision medicine, genome assembly

Since the landmark completion of the Human Genome Project, the biomedical field has relied primarily on single linear reference genomes, such as the widely used GRCh38, as the standard template for genetic studies. While invaluable, these singular consensus sequences are inherently limited by their inability to capture the breadth of human genetic diversity. Human populations exhibit immense variability, with complex genetic variations including structural variants (SVs) and tandem repeats (TRs) often escaping detection within traditional single-reference frameworks. Recognizing these limitations has driven researchers toward the conceptual framework of a pangenome—a comprehensive genomic repository that embodies the collective genetic landscape of a population or species, thereby encompassing variants that are rare or absent in standard references.

Despite advances in long-read sequencing technologies that enable high-fidelity diploid genome assemblies, the prohibitive costs have historically restricted pangenomic sampling to narrow cohorts numbering only a few dozen individuals. Such small datasets lack sufficient statistical power to accurately discern variant frequencies or resolve variants existing at low population frequencies and within highly complex genomic regions. The urgent need for scalable and cost-efficient sequencing methodologies has become a critical bottleneck in the large-scale exploration and functional characterization of genomic diversity.

Addressing this challenge, Jian Yang’s research group, renowned for its methodological innovations in statistical genetics, computational genomics, and big data analytics related to human complex traits, has devised the Pangenome-Informed Genome Assembly (PIGA) workflow. Unlike traditional de novo assembly methods that solely depend on sequencing data from individual genomes, PIGA implements a pangenome-guided framework that integrates sequencing information holistically across the entire cohort. This approach capitalizes on a hybrid sequencing strategy combining moderate-coverage Illumina short reads with PacBio long reads, effectively reducing sequencing expenses without compromising assembly quality. Such a balanced strategy enables the assembly of high-quality genomes from modest-coverage data, presenting a practical and scalable path for future population-scale hybrid sequencing projects.

Applying the PIGA workflow, the researchers constructed the largest human pangenome published to date—the 1,000 Chinese Pangenome—encompassing 1,116 diploid assemblies with an impressive mean quality value (QV) of 46. This robust collection yielded an astonishing 405.3 million base pairs of non-reference genomic sequences absent from current leading human references such as GRCh38 and the newer CHM13 assembly. Remarkably, within this non-reference fraction, the team annotated 26.2 million base pairs as functional genic elements and predicted regulatory regions. This unprecedented expansion into the territory of non-reference genome sequences significantly enhances our understanding of genomic complexity beyond canonical references.

Harnessing such a comprehensive and high-quality dataset, the team generated an exhaustive catalog of genomic variation spanning multiple variant types and scales. This included more than 35 million small variants alongside a vast repertoire of complex structural variants (110,530 SVs), tandem repeats (485,575 TRs), and nearly 860,000 nested variants embedded within non-reference sequences. This rich catalog provides an invaluable resource for dissecting genomic variation with greater resolution than previously possible, capturing the multifaceted nature of human genomic architecture.

Of particular clinical importance, the researchers mapped medically relevant variations at diverse scales, illustrating how the 1,000 Chinese Pangenome (1KCP) variant catalog serves as a critical reference for clinical genetics. Noteworthy findings included gene-altering structural variants that disrupt coding sequences, expansions of pathogenic tandem repeats implicated in disease, variable gene clusters, and highly polymorphic human leukocyte antigen (HLA) gene haplotypes that affect immune function. Collectively, these insights herald significant improvements in the screening and understanding of pathogenic mutations across populations.

Complementing the genomic identification of variants, the study integrated extensive gene expression data to perform pan-variant expression quantitative trait loci (eQTL) mapping. This innovative analysis uncovered 3,256 eQTLs linked to complex variant types—specifically structural variants, tandem repeats, and nested variants—highlighting the elaborate regulatory interplay between diverse forms of genetic variation and gene expression. These findings underscore the complexity embedded within regulatory landscapes sculpted by intricate genomic architectures.

The implications of this work extend broadly beyond the immediate findings. By pioneering a cost-effective hybrid sequencing approach married to a cohort-wide pangenome-guided framework, this study charts a transformative course for future genomic investigations across human populations and other species alike. The ability to characterize complex genomic variation comprehensively—and to functionally elucidate its biological impact—opens new frontiers in precision medicine, evolutionary biology, and genetic epidemiology.

Ph.D. student Yifei Wang and Research Assistant Professor Zhongqu Duan share co-first authorship on the study, with Professor Jian Yang serving as the senior author. The ambitious project received financial support from key national funding bodies, including the National Natural Science Foundation of China, the National Key R&D Program, and Zhejiang’s “Pioneer & Leading Goose” Program, alongside the New Cornerstone Science Foundation. High-performance computation for the analyses was facilitated by the Westlake University High-Performance Computing Center.

Professor Jian Yang’s research group continues to push the boundaries of statistical genetics and bioinformatics method development, probing the genetic underpinnings of complex diseases through large-scale multi-omic population datasets. Their mission is to unravel the intricate genetic architecture governing disease traits, developing translational approaches that advance diagnostic accuracy, therapeutic target discovery, and tailored precision medicine strategies for improved human health.

This landmark publication not only sets a new standard for human pangenome assembly but also heralds a paradigm shift, emphatically demonstrating that integrating cost-effective hybrid sequencing with innovative computational frameworks can unlock the vast genomic diversity requisite for cutting-edge biomedical discoveries in the genomic era.

Subject of Research: Not applicable

Article Title: The 1000 Chinese Pangenome empowers medical and population genetics

News Publication Date: 1-Apr-2026

References:
DOI: 10.1038/s41586-026-10315-y

Image Credits: Jian Yang Lab at Westlake University

Keywords: Computational biology, genome assembly, pangenome, structural variants, tandem repeats, population genetics, hybrid sequencing, genomic diversity, precision medicine