Multidimensional variation and population stratification across 8000 complete human centromeres
Yanqing SuniD, Shijie WaniD, Lei Nie, Dan YuiD, Feifei Zhou, Yi YangiD, Xiangyu YangiD, An LiuiD, Quanyu CheniD, Kezhi Fu, Qingyang NiiD, Yaoxi HeiD, Bing SuiD, Yafei MaoiD, Kai YeiD, Xiaofei Yang, Guojie ZhangiD, Wu Dongya
Abstract
Human centromeres are indispensable for the faithful segregation of chromosomes during cell division, yet their highly repetitive nature has historically precluded comprehensive characterization, leaving fundamental questions about their sequence diversity, evolution trajectories and function dynamics unresolved. Here, we generated 6,312 complete human centromere sequences from 320 phased genome assemblies in Asian Pan-Genome project phase 1. By integrating the assemblies from the Human Pangenome Reference Consortium (HPRC) and Human Genome Structural Variation Consortium (HGSVC), we constructed a multidimensional genetic variation map encompassing over 8,000 gapless centromeres. Centromeric satellite arrays account for 4.19% to 6.01% of the whole genome, with substantial variations in size and architecture across chromosomes. Using a refined alpha satellite clustering approach that captures global diversity, we identified 195 higher-order repeat (HOR) arrays, 56.4% of which are absent from the T2T-CHM13 reference genome. Extensive structural variations across multiple dimensions exhibit population stratification, including centromeric haplotypes (CenHaps), ultra-large pericentric inversions spanning up to 36.6 Mbp, and inter-chromosomal HOR sharing that reflects sequence exchange among chromosomes. Integrating CENP-A CUT&Tag experiments and long-read-based DNA methylation profiles, we demonstrate that 16.8% of centromeres harbor multiple potential kinetochore assembly sites, and CenHap-specific local HOR homogenization is associated with kinetochore positioning. Despite global suppression of recombination at centromeres, we observed asymmetric linkage disequilibrium flanking centromeres and an ancient recombination event within the centromere of chromosome 19. Furthermore, contrary to the prevailing assumption of high mutation rates in centromeres, our estimates based on stringent orthology reveal no significantly higher single-base substitution rates for centromeres relative to flanking pericentromeric regions with substantial variations across chromosomes, despite extraordinary structural plasticity. Collectively, these multi-scale centromeric variations provide a global view of human centromere diversity and population stratification, fundamentally redefine centromere evolution through a dual-track model balancing structural innovation with mutational constraint, and establish an essential resource for investigating centromere biology and a baseline reference for diagnosing centromere-associated disorders.
02 /Specification & Classification
Click any tag to browse matching papers- Citations
- 2
- Organism
- Assay
- Research Focus
- Approach
In-depth summary
Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 1 (24 Jul 2026) · CC0 1.0
Over 8,000 gapless human centromeres reveal population-stratified satellite-array, inversion and haplotype variation, multiple candidate kinetochore sites in 16.8% of centromeres, and single-base substitution rates comparable to flanking pericentromeres.
Earlier pangenomes completed fewer centromeres (20.6% gapless in HPRC year 1, 66.3% in HGSVC phase 3), and hidden-Markov-model alpha satellite annotation built on limited diversity missed population-specific variants. Here 85.8% of centromeres in 320 East Asian haplotype assemblies are gapless, and reference-free monomer clustering plus a collinearity-aware aligner (HORSCAN) allow population-level comparison of higher-order repeat (HOR) structure, kinetochore sites and substitution rates.
Key results
Shown6,312 of 7,360 centromeres (85.8%) in 320 East Asian haplotype assemblies were assembled gapless. Size differences between maternal and paternal centromeres mostly fall in a narrow band, which the authors propose as a reference range for meiotic risk.
Gapless rate 85.8% vs 20.6% (HPRCy1) and 66.3% (HGSVC3); average centromere QV 85.9, checked with GCI, NucFreq, Flagger, GAVISUNK and VerityMap. In 160 individuals, 94.3% (2,581 of 2,736) of parental centromere pairs fell within mean ± 2 s.d. (-4.54 to 4.53 Mbp); nine pairs differed by >10 Mbp, eight of them on chromosomes 1, 9 and 16. · A comprehensive atlas of human complete…; Evaluating centromere assembly quality
ShownClustering alpha satellite (αSat) monomers from all cohorts without a reference adds many HOR arrays missing from T2T-CHM13. It also shows HOR arrays shared between chromosomes, mostly within two centromere groups.
~244 million monomers grouped into 29,773 clusters; 3,722 form HORs (95.8% of αSat sequences), 2,197 (59.0%) of them absent from T2T-CHM13; 57 of 60 known CHM13 HOR arrays recalled. HiCAT called 195 HOR arrays, 56.4% absent from T2T-CHM13 (110 novel). 25 arrays are shared by at least two chromosomes, mainly CEN13/14/21/22 (n = 15) and CEN01/05/16/19 (n = 8). · HOR diversity of human alpha satellites; Inter-chromosomal HOR sharing landscape; Global monomer clustering algorithm; De novo HOR inference
ShownThe frequencies of large inversions and of centromere haplotypes (CenHaps, defined by HORSCAN synteny distance) differ by ancestry. A pericentric inversion that spans the whole chromosome 9 centromere passes through meiosis despite its size.
194 large (>10 Kbp) peri/centromeric inversions across 538 haplotype assemblies. A ~36.6 Mbp CEN09 inversion was heterozygous in ten individuals (six East Asian, four African; 5/349 unrelated East Asian, 3/97 African, none in other superpopulations) and passed from father to offspring in two families. CEN08 CenHap-C: 20 assemblies, 19 African; CEN04 CenHaps diverged ~321-553 thousand years ago. · Large peri/centromeric inversions; Population stratification of centromere…; Inversion validation; Phylogenetic inference and CenHap divergence…
SuggestedHypomethylated centromere dip regions (CDRs) mostly match CENP-A CUT&Tag peaks, and 16.8% of centromeres carry more than one CDR. Where the CDR sits differs between CenHaps, linking structural haplotype to kinetochore placement.
CENP-A CUT&Tag peaks (induced pluripotent stem cells from five individuals) agreed with blood-methylation CDRs in 165 of 183 centromeres (90.2%); additional CDRs (>80 Kbp from primary) on 1,482 chromosomes (16.8%); 99.94% of CDRs lay in active HOR arrays, with higher local identity than whole active arrays (p = 9.12×10^-97). In CEN08, 94.4% of CenHap-A CDRs sat outside recently expanded zones, unlike CenHap-B. · Epigenetic dynamics among centromeres; CUT&Tag library preparation and sequencing; CUT&Tag data analysis; CDR analysis
ShownOnce orthology is strictly filtered, single-base substitution frequencies in active αSat HOR arrays are no higher than in flanking pericentromeric sequence, despite heavy structural turnover. Rates vary widely between chromosomes.
93 orthologous centromere pairs (>99% identity, paralogous segments masked): median centromere-to-pericentromere substitution ratio 1.027, p = 0.70 (Wilcoxon signed-rank); the same held in 64 parent-offspring pairs. Rates varied over 100-fold across chromosomes; 14 pairs had no substitutions. CDRs looked lower than non-CDR regions in 20 pairs (16 with zero CDR substitutions), flagged as tentative. · Centromeric recombination and single-base de…; Comparing centromeric and pericentromeric…
Methods, data and reproducibility
Phased HiFi plus Oxford Nanopore (ONT) ultra-long assemblies of 160 East Asian individuals were pooled with HPRCy1, HGSVC3 and four reference genomes. Centromeres were annotated with reference-free monomer clustering, HORmon/HiCAT and HORSCAN. Kinetochore sites came from CENP-A CUT&Tag and CpG methylation; linkage and substitution rates from graph SNPs and curated orthologous pairs.
- Asian Pan-Genome project phase 1 (APGp1) phased haplotype assemblies (PacBio HiFi ~51×, ONT ultra-long ~62×; trio and Hi-C modes), East Asian individuals · human · n = 160 individuals; 320 haplotype assemblies
- Human Pangenome Reference Consortium year 1 (HPRCy1) assemblies, scaffolded with RagTag · human · n = 43 (excluding HG002 and individuals overlapping HGSVC3)
- Human Genome Structural Variation Consortium phase 3 (HGSVC3) assemblies, scaffolded with RagTag · human · n = 65
- Reference-level complete assemblies: T2T-CHM13, T2T-CN1, Q100-HG002 and YAO · human · n = 4
- CENP-A CUT&Tag with IgG controls on PBMC-derived induced pluripotent stem cells · human · n = 5 individuals · NGDC HRA014419 (HRX2355018 to HRX2355022)
- ONT methylation from lymphoblastoid cell line samples used to compare CDR positions across cohorts · human · n = 20 HGSVC3 and 10 HPRCy1 samples
- Chimpanzee and bonobo sequences as dating outgroups; gorilla CEN16 region for HOR synteny comparison · chimpanzee, bonobo, gorilla
- Methods & tools
- hifiasm, Verkko, VSEARCH, HORmon, HiCAT, HORSCAN, CDR-Finder, PAV, LSGvar, PLINK, IQ-TREE, BEAST
- Code
- https://github.com/Asian-Pan-Genome/Centromerehttps://github.com/XDwan/HORSCANhttps://github.com/Hanjunmin/LSGvar
- Data & accessions
- https://github.com/Asian-Pan-Genome/Centromerehttps://genome.zju.edu.cn/APG/cenview/https://doi.org/10.5281/zenodo.17647353https://genome.zju.edu.cn/APG/download/https://ngdc.cncb.ac.cn/
- Software & environment
- Tool versions are given, e.g. minimap2 v2.26-r1175, Winnowmap2 v2.03, Flagger v0.3.2, VerityMap v2.1.2, Merqury v1.4.1, RepeatMasker v4.1.2 (Dfam v3.3), VSEARCH v2.30.0, PAV v2.4.6, IQ-TREE v2.3.6, BEAST v2.7.7, PLINK v1.9.0-b.7.7; no container or requirements file is described.
- Reproducibility gaps
- CENP-A CUT&Tag reads are under controlled access (NGDC HRA014419; request to the DAC via NGDC). No accession is printed for the APGp1 assemblies or raw reads. Several steps (inversion calls, CDR calls, choice of substitution pairs) rely on manual curation.
Limitations (6)
- CENP-A enrichment is underestimated for some chromosomes (e.g., chromosome Y) because short CUT&Tag reads cannot be placed uniquely in near-identical centromeric sequence. · stated by the authors
- CenHaps built on 96%-identity monomer clusters lack resolution in highly homologous arrays. There, synteny distance is driven by alignment gaps and array size, so the groupings may not reflect phylogeny and need caution. · stated by the authors
- Relative substitution rates vary more than 100-fold across chromosomes, and the CDR versus non-CDR comparison rests on few pairs and short CDRs. Pedigree assemblies are needed for intergenerational rates. · stated by the authors
- Several questions stay open: whether ancestry stratification reflects drift or selection, why linkage is asymmetric across the centromere, and where exactly centromeric recombination breakpoints lie. βSat bimodality on acrocentrics may reflect residual assembly incompleteness. · stated by the authors
- Kinetochore positions rest mostly on methylation dips measured in blood. Direct CENP-A data cover stem cells from only five individuals, so the link from CenHap to kinetochore placement is an association, not a tested mechanism. · noted in this summary
- All non-East Asian diversity comes from HPRCy1 (43) and HGSVC3 (65) assemblies, so ancestry-specific frequencies rest on small counts (e.g., 3/97 African carriers of the CEN09 inversion). · noted in this summary
Who should read:Read this if you work on centromere biology, satellite DNA annotation, pangenome structural variation or kinetochore epigenetics and want population-scale complete centromere annotations and HOR analysis tools.
Reproducibility ledger
Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.
Sign in to file a reproducibility report.
Related papers
- The variation and evolution of complete human centromeres
Glennis A. Logsdon, Allison N. Rozanski, … Evan E. Eichler·NaturePeer-reviewed·Apr 3, 2024
- A global view of human centromere variation and evolution
Shenghan Gao, Keisuke K. Oshima, … Pi-Chuan Chang·NaturePeer-reviewed·Jul 29, 2026
- Centromere landscapes resolved from hundreds of human genomes
Shenghan Gao, Yimeng Zhang, … Kai Ye·Genomics Proteomics & BioinformaticsPeer-reviewed·Oct 1, 2024
- Pangenome alignment reveals global diversity and evolution of human centromeric regions
Jordan M. Eizenga, Mira Mastoras, … Karen H. Miga·bioRxiv·Sep 4, 2026
- Complete genomic and epigenetic maps of human centromeres
Nicolas Altemose, Glennis A. Logsdon, … Karen H. Miga·SciencePeer-reviewed·Mar 31, 2022
- Long-read Data Revealed Structural Diversity in Human Centromere Sequences
Yuta Suzuki, Eugene Wimberly Myers, Shinichi Morishita·bioRxiv·Sep 27, 2019