A telomere-to-telomere gap-free genome assembly of apricot (Prunus armeniaca L.) Authors: Zhenyu Huang, Hangling Bie, Xinyue Su, Lehan Xia, Long Chen, Yuling Chen, Ruigang Wu, Ming Li, Pingyong Wang, Ke Cao Published: 2026 ABSTRACT AbstractApricot (Prunus armeniaca L.), a globally cultivated temperate fruit crop, represents an economically critical species for fresh consumption and value-added products. Despite its agricultural importance, genomic resources have lagged behind other Prunus species, hampering trait dissection and molecular breeding. Here, we present the first telomere-to-telomere (T2T) gap-free genome assembly for the elite cultivar ‘Sungold’, achieved through the integration of multi-platform sequencing technologies: Illumina short-read, PacBio HiFi, ONT ultra-long, and Hi-C scaffolding. The final assembly spans 251.36 Mb with benchmark quality metrics: a contig N50 of 32.04 Mb, BUSCO completeness score of 98.8%, LTR Assembly Index (LAI) of 15.99, and a consensus quality value (QV) of 59.75. The T2T assembly achieved the resolution of all eight centromeres, together with fourteen intact telomeres, confirming its high structural integrity. Genomic annotation revealed 43.38% of repetitive sequences and 25,999 predicted protein-coding genes. This gap-free T2T genome resource establishes the highest-resolution reference for investigating apricot genome evolution, structural variation, and trait-associated genetic mechanisms in modern breeding programs.Subject terms: Plant sciences, Genome Background & Summary Apricot (Prunus armeniaca L.), a key member of the genus Prunus (subgenus Prunophora) within the Rosaceae family, stands as one of the world’s most economically significant temperate fruit crops1. Global production exceeds 4.0 million tonnes annually2, underpinning extensive cultivation driven by the fruit’s exceptional versatility. Apricots are highly prized by consumers worldwide for their distinctive, delicious flavor, captivating aroma, and considerable nutritional value, consumed fresh, dried, canned, and processed into various products. In China, apricots hold profound cultural significance, celebrated not only for their superb taste but also for their exquisite, early-spring ornamental blossoms that have inspired poets and artists for centuries. Renowned for its excellent fruit quality and superior overall horticultural traits, the ‘Sungold’ apricot stands as one of the world’s predominant commercial apricot cultivars. Its exceptional characteristics, including high yield potential, market adaptability, and consistent performance, have cemented its status as a backbone breeding parent in China’s apricot breeding programs, where it has been instrumental in developing numerous successful new varieties. Understanding the genetic underpinnings of these valuable traits of P. armeniaca necessitates high-quality genomic resources. To date, contig-level genome assemblies of several apricot cultivars have been generated, including ‘Sungold’, ‘Chuanzhihong’, ‘Yinxiangbai’, ‘Xiaobaixing’, and others3–9. These draft assemblies have provided foundational insights for comparative and evolutionary genomic studies within Prunus and Rosaceae. However, the presence of persistent gaps and unresolved repetitive regions in these assemblies has left certain complex genomic structures, such as telomeres and centromeres, incompletely characterized. The completion of a truly complete, telomere-to-telomere (T2T), gap-free genome assembly represents a major pinnacle in genomics research. Fueled by revolutionary advances in long-read sequencing technologies—specifically PacBio High-Fidelity (HiFi) sequencing offering exceptional accuracy, Oxford Nanopore Technology (ONT) ultra-long reads spanning hundreds of kilobases to resolve complex repeats, and High-throughput chromosome conformation capture (Hi-C) for accurate scaffolding—T2T assemblies have recently been achieved for several major fruit trees. These landmark references include kiwifruit10, banana11, grapevine12, apple13,14, pear15, peach16, and plum17. Such T2T genomes are transformative, enabling the precise identification and characterization of genes, structural variations (SVs), and regulatory elements within previously inaccessible genomic “dark matter”, such as centromeres, telomeres, and dense TE clusters, crucial for understanding genome evolution, stability, and trait regulation18. Addressing the critical gap in apricot genomics and leveraging these technological breakthroughs, we present the first T2T gap-free genome assembly for the elite cultivar ‘Sungold’. This high-resolution assembly was meticulously constructed through the strategic integration of multi-platform sequencing data: Illumina short-reads for high-accuracy base correction, PacBio HiFi long-reads for precise assembly of genic regions, ONT ultra-long reads to span and resolve complex repetitive structures (including centromeres and telomeres), and Hi-C data for accurate chromosome-level scaffolding. This unprecedented, gap-free T2T genome assembly of ‘Sungold’ apricot establishes the highest-resolution reference available for the species. It provides an invaluable, foundational resource poised to revolutionize research into apricot genome evolution, the discovery and functional characterization of SVs and genes underlying key agronomic traits (e.g., fruit quality, disease resistance, flowering time, adaptability), and ultimately, to accelerate precision breeding efforts for the development of superior apricot varieties. Methods Sample collection, DNA and RNA extractions Fresh leaves, young fruits, bark from tender branches, and seeds were collected from a five-year-old ‘Sungold’ apricot tree cultivated at the experimental orchard of the Zhengzhou Fruit Research Institute, Chinese Academy of Agricultural Sciences (34°42′36″N, 113°42′25″E) during the active growth phase. All collected samples were immediately snap-frozen in liquid nitrogen and subsequently transferred to long-term storage at −80 °C for preservation of biomolecular integrity. Genomic DNA (gDNA) intended for Illumina short-read and PacBio HiFi sequencing was isolated using the DNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA) in strict accordance with the manufacturer’s recommended protocols. For ONT ultra-long sequencing, high-molecular-weight gDNA was extracted via an optimized sodium dodecyl sulfate (SDS)-based extraction. Parallel to DNA extraction, total RNA from all tissue types was isolated using the RNeasy Plus Mini Kit (Qiagen, Valencia, CA, USA) following the manufacturer’s instructions. Comprehensive quality assessment of nucleic acids was performed through dual approaches: spectrophotometric analysis using a NanoDrop 2000 instrument (NanoDrop Technologies, Wilmington, DE, USA) and electrophoretic evaluation via agarose gel electrophoresis. DNA quantity was determined using the Qubit dsDNA HS Assay Kit on a Qubit 3.0 Fluorometer (Life Technologies, Carlsbad, CA, USA), while RNA concentration and integrity were determined using the Bioanalyzer 2100 system (Agilent Technologies, CA, USA). Library construction and sequencing The Illumina sequencing workflow involved preparation of a 350 bp insert-size library following the Illumina TruSeq DNA PCR-Free Library Prep Kit protocol, which was subsequently sequenced on an Illumina NovaSeq. 6000 platform (Illumina, San Diego, CA, USA) to generate 150 bp paired-end (PE150) reads. PacBio HiFi sequencing library was constructed using the SMRTbell Express Template Prep Kit 2.0 according to the manufacturer’s recommendations, with final sequencing conducted on the PacBio Sequel II system (PacBio Biosciences, CA, USA). For ONT ultra-long sequencing, the library was prepared with the Oxford Nanopore SQK-LSK110 kit following the manufacturer’s instructions and sequenced using a PromethION sequencer (Oxford Nanopore Technologies, UK). Hi-C libraries were generated through an established chromatin conformation capture methodology as previously described13, involving formaldehyde treatment, restriction enzyme digestion, biotin labeling, ligation with connector enzymes, and DNA capture with labeled probes; these libraries were then sequenced on the Illumina NovaSeq. 6000 platform (PE150) (Illumina, San Diego, CA, USA). Transcriptomic analysis employed complementary approaches: tissue-specific short-read RNA-seq libraries for leaves, fruits, bark, and seeds were individually prepared and sequenced on the Illumina NovaSeq. 6000 (PE150), while full-length transcriptome characterization was achieved by constructing a pooled cDNA library from all four tissues followed by sequencing on the PacBio Sequel II system using the kinnex Iso-Seq single-molecule long-read workflow (PacBio Biosciences, CA, USA). Cumulatively, this integrated sequencing strategy generated the following clean datasets: 33.27 Gb of Illumina short reads (~116 × ), 142.05 Gb of PacBio HiFi reads (~493 × with read N50 = 18.29 kb), 39.94 Gb of ONT ultra-long reads (~139 × with read N50 = 100 kb), 39.77 Gb of Hi-C paired-end data (~138 × ), tissue-specific short-read RNA-seq data (6.1 Gb, ~21 × from fruit; 6.0 Gb, ~21 × from bark; 7.0 Gb, ~24 × from leaf; and 6.2 Gb, ~22 × from seed), and 9.23 Gb (~32 × ) of PacBio Iso-Seq long-read transcriptomic data. All of these datasets were utilized for subsequent genome assembly and annotation. Genome survey Prior to genome assembly, we conducted a K-mer analysis with Jellyfish (v2.2.7)19 and GenomeScope (v2.0)20 to assess genomic characteristics of the ‘Sungold’ apricot. Analysis of the 21-mer frequency distribution revealed an estimated genome size of 287.93 Mb, a heterozygosity rate of 1.11%, a GC content of 38.8%, and a repeat content of 48.82% (Fig. 1, Table 1). Frequency distribution of the 21-mer analysis. The x-axis represents the K-mer depth and y-axis represents the frequency of K-mer corresponding to the depth. Table 1Statistics of the Prunus armeniaca L. genome assembly.Genomic featuresValueEstimated genome size (Mb)287.93Estimated heterozygosity rate (%)1.11Assembled genome size (Mb)251.36Contig N50 length (Mb)32.04Number of base chromosomes8Number of gap-free chromosomes8Number of candidate telomeres14Number of candidate centromeres8Repetitive sequences (%)43.38Quality value (QV)59.75LTR assembly index (LAI)15.99Genome BUSCOs (%)98.8Number of genes25,999Gene BUSCOs (%)98.3 Statistics of the Prunus armeniaca L. genome assembly. Genome assembly The draft genome assembly of ‘Sungold’ was initially generated using Hifiasm (v0.20.0)21, yielding 402 contigs with an N50 of 32.04 Mb and a total assembly size of 283.46 Mb. The contig-level assembly was subsequently ordered, oriented, and anchored into pseudo-chromosomes using YaHS (v1.1)22 with Hi-C interaction data, resulting in a chromosomal level assembly. Manual adjustment and visualization was performed based on chromosomal interaction intensity using Juicebox (v1.11.08)23 and HapHiC (v1.0.5)24, respectively (Fig. 2). Aiming for T2T continuity, all remaining gaps were resolved using ONT ultra-long reads processed by quarTeT (v1.1.7)25. To enhance base accuracy, the assembly underwent error correction with HiFi data using Inspector (v1.0.2)26. Finally, the ‘Sungold’ genome assembly achieved a gap-free configuration with a refined total size of 251.36 Mb and a contig N50 of 32.04 Mb, exhibiting a significant improvement compared to the previously reported apricot genomes (Tables 1, 2). The genomic sequences were clustered without gaps and localized on eight chromosomes (Fig. 3). Heterozygosity (1.11%), repetitive sequences (duplication rate = 1.15), and organellar DNA are likely responsible for the larger genome size estimated by K-mer analysis compared to the final assembly. Hi-C intra-chromosomal interaction heatmap of the Prunus armeniaca L. genome. Table 2Statistics for the genome assemblies of Prunus armeniaca L.Apricot cultivarsTotal length (Mb)Contig N50 (Mb)Scaffold N50 (Mb)Genome BUSCOsReferencesSungold251.3632.0432.0498.8%this studyChuanzhihong221.901.0225.1398.0%ref. 4Rojo Pasión Hap1216.000.6625.8097.9%ref. 5Rojo Pasión Hap2215.200.6625.5097.9%ref. 5Marouch #14203.931.7925.1597.3%ref. 6Stella212.061.3125.5297.8%ref. 6Yinxiangbai251.333.1730.9896.2%ref. 8Xiaobaixing231.893.5626.7396.2%ref. 9Note: ref. 4 refers to the fourth reference in the reference list of the manuscript, and so forth for the others. Statistics for the genome assemblies of Prunus armeniaca L. Note: ref. 4 refers to the fourth reference in the reference list of the manuscript, and so forth for the others. Overview of the Prunus armeniaca L. genome assembly. The tracks from outside to inside are: chromosome ID and length (I), gene density (II), repeat density (III), ncRNA content (IV), and GC content (V). The innermost track displays a photograph of the ‘Sungold’ apricot. Repetitive sequences annotation Repetitive elements were annotated using a combination of de novo prediction and homology-based approaches. De novo repeat prediction was performed using RepeatModeler (v1.0.11)27, supplemented by LTR-FINDER (v1.0.7)28 for specific detection of long terminal repeats (LTRs). For homology-based annotation, RepeatMasker (v4.0.9)29 and RepeatProteinMasker (v4.0.9)29 were employed to detect DNA and protein transposable elements, respectively, against the Repbase database30. Tandem Repeat Finder (v4.10.0)31 was used to identify tandem repeats. This comprehensive analysis revealed 109.04 Mb of the ‘Sungold’ genome as repetitive sequences, representing 43.38% of the total genome. LTR retrotransposons dominated the repetitive landscape, comprising 57.30% of all repetitive elements (Table 3). Additionally, centromeric sequences were identified using quarTeT (v1.1.7)25, and telomeric regions (CCCATTT at the 5’ end and TTTAGGG at the 3’ end) were identified using tidk (v0.2.41)32. This analysis revealed a nearly complete chromosomal structure, including eight centromeres and fourteen telomeres, confirming the high continuity and integrity of the genome assembly (Fig. 4, Tables 4, 5). Table 3Summary of the repetitive sequences identified in the Prunus armeniaca L. genome.ClassSubclassCountMasked (bp)Masked (%)Class ICopia13,0238,383,5203.34Gypsy15,73743,847,27417.44LTR_unknown23,58910,253,7234.08LINE_element1,150715,8980.28nonLTR_unknown16693,7920.04Class IICACTA13,1446,333,0942.52Mutator35,15410,207,3594.06PIF_Harbinger16,6825,747,6492.29Tc1_Mariner1,368355,2680.14hAT11,0343,911,9761.56helitron15,1693,835,4241.53Others38,59515,355,5966.11Total184,811109,040,57343.38 Summary of the repetitive sequences identified in the Prunus armeniaca L. genome. Localization of telomeres and centromeres in the Prunus armeniaca L. genome. Triangles and circles represent telomeres and centromere, respectively; red indicates high gene density while blue indicates low gene density. Table 4Positions of the candidate telomeres of the Prunus armeniaca L. genome.ChromosomesLength (bp)HeadTailStart (bp)End (bp)Start (bp)End (bp)Chr0149,405,31942,57549,403,13649,405,313Chr0232,192,20721,57332,190,15432,192,206Chr0332,855,67442,18432,853,58032,855,674Chr0429,753,733192,297NANAChr0522,791,67881,30622,789,02322,791,672Chr0632,040,05741,58532,038,56432,040,057Chr0727,482,570NANA27,480,29227,482,567Chr0824,834,30972,25124,832,71724,834,303Note: NA, telomeric repeats not detected due to assembly gaps, sequence absence, or detection thresholds. Positions of the candidate telomeres of the Prunus armeniaca L. genome. Note: NA, telomeric repeats not detected due to assembly gaps, sequence absence, or detection thresholds. Table 5Positions of the candidate centromeres of the Prunus armeniaca L. genome.ChromosomesStart (bp)End (bp)Length (bp)Chr0120,332,41824,091,3113,758,893Chr029,228,56012,576,2323,347,672Chr0311,556,30118,200,7996,644,498Chr0421,535,94825,735,7004,199,752Chr058,400,00012,900,0014,500,001Chr0615,599,37318,474,3402,874,967Chr074,400,0009,200,0004,800,000Chr087,725,97710,216,2022,490,225 Positions of the candidate centromeres of the Prunus armeniaca L. genome. Prediction and functional annotation of protein-coding genes Protein-coding genes in the ‘Sungold’ T2T genome were predicted through a combination of three complementary approaches: ab initio prediction using AUGUSTUS (v3.4.0)33; homology-based prediction using GeneWise (v2.4.1)34; and transcriptome-assisted gene prediction using HISAT2 (v2.1.0)35 and PASA (v2.4.1)36. First, the gene structures predicted by homology-based and transcriptome-assisted methods were integrated using EviAnn (v2.0.2)37, generating 23,964 evidence-supported genes. To additionally capture gene models not represented in the evidence-based integration, we supplemented de novo genes predicted by AUGUSTUS that met the following criteria: (i) supported by expression evidence (RNA-seq), and (ii) showing no genomic overlap with existing EviAnn-annotated genes. We trained ab initio parameters for AUGUSTUS using a high-confidence training set derived from our RNA-seq data to optimize performance. This supplementation resulted in a final high-confidence set of 25,999 protein-coding genes. Functional annotation of the predicted proteins was carried out via BLASTP (v2.6.0)38,39 searches against multiple databases. These included the Kyoto Encyclopedia of Genes and Genomes (KEGG)40, Gene Ontology (GO)41, Clusters of Orthologous Groups of proteins (COGs)42, evolutionary genealogy of genes: Non-supervised Orthologous Groups (eggNOG)43, and the protein families (Pfam) database44. The final annotation comprised 25,999 protein-coding genes. These genes exhibited an average mRNA length of 3,607.7 bp and contained 5.3 exons per gene. The average coding sequence (CDS) length was 1,263.7 bp, with an average exon length of 309.5 bp. Notably, 94.6% (24,596 genes) of the predicted genes received functional annotation in at least one queried database, highlighting the high completeness and reliability of the ‘Sungold’ T2T genome annotation (Tables 6, 7). Table 6Summary of the predicted genes in the Prunus armeniaca L. genome.Gene setNumberAverage gene length (bp)Average CDS length (bp)Average exon per geneAverage exon length (bp)de novo/AUGUSTUS26,8193,292.441,219.485.41297.18homo/P.persica29,9343,138.881,211.394.69258.34homo/P.mume38,4463,154.051,163.184.61252.26homo/P.avium44,4852,860.451,010.393.87260.95homo/P.salicina28,8103,180.761,184.605.04235.15homo/P.dulcis43,9212,675.711,096.373.86283.87trnascript/RNA_seq21,9423,454.561,195.285.23367.21trnascript/isoform43,4145,644.811,104.106.38567.73EviAnn23,9643,718.601,284.965.34315.94Final25,9993,607.661,263.735.27309.51 Summary of the predicted genes in the Prunus armeniaca L. genome. Table 7Annotation of the protein-coding genes in the Prunus armeniaca L. genome.TypeNumberPercentage (%)COG22,64387.09eggNOG24,59694.6GO12,17946.84KEGG11,55044.42Pfam22,07384.9Annotated gene24,59694.6Total gene25,999100 Annotation of the protein-coding genes in the Prunus armeniaca L. genome. Non-coding RNA annotation Transfer RNAs (tRNAs) and ribosomal RNAs (rRNAs) were identified using tRNAscan-SE (v2.0.7)45 and Barrnap (v0.9)46, respectively. Additional non-coding RNA (ncRNA) types, including microRNAs (miRNAs), and small nuclear RNAs (snRNAs), were annotated using Infernal (v1.1.3)47 with searches against the Rfam database (v14.2)48. Comprehensive annotation revealed a total of 4,371 ncRNAs in the ‘Sungold’ T2T genome, comprising 3,436 rRNAs, 517 tRNAs, 263 miRNAs, and 155 snRNAs (Table 8). Table 8Annotation of non-coding RNA genes in the Prunus armeniaca L. genome.ClassTypeCopyAverage length (bp)Total length (bp)% of genomemiRNAmiRNA263112.2529,5210.01174tRNAtRNA51775.7739,1740.01559rRNA18S3501,723.23603,1290.2399528S1,325142.15188,3470.074935.8S333161.5253,7860.02145S1,072124.65133,6200.05316CD-box356102.4936,4850.01452snRNAHACA-box37128.924,7700.0019splicing118145.9217,2180.00685 Annotation of non-coding RNA genes in the Prunus armeniaca L. genome. Comparative genomic analysis To identify structural variants between the contig-level assembly and the newly completed T2T assembly of the ‘Sungold’ apricot genome, we performed a comparative analysis using SyRI (v1.6.3)49 with default parameters (Fig. 5). The analysis revealed substantial improvements in genome continuity, including the closure of 172 residual gaps (totaling 1.27 Mb) present in the contig-level assembly. Additionally, SyRI detected 894 translocations (2.55 Mb, ~1.01% of the genome), 51 inversions (1.05 Mb, ~0.42%), and 106 duplications (0.18 Mb, ~0.07%). Comparative alignment assessments confirmed significant enhancements in chromosomal anchoring and orientation accuracy in the T2T assembly compared to the previous version. Sequence collinearity and structural variants between the contig-level and the novel telomere-to-telomere assembly of ‘Sungold’ apricot. Data Records The sequencing datasets, which include Illumina reads, PacBio HiFi reads, ONT ultra-long reads, Hi-C data, and transcriptomic reads, have been deposited at the Genome Sequence Archive (GSA) in National Genomics Data Center (NGDC) under the accession CRA02765650, and have also been deposited at the Sequence Read Archive (SRA) in National Center for Biotechnology Information (NCBI) under the accession SRP60284051. The T2T genome assembly of ‘Sungold’ has been deposited at the NGDC Genome Warehouse (GWH) under the accession GWHIGUE00000000.152 and at the NCBI GenBank under the accession JBWBED00000000053. Additionally, the T2T genome assembly and its corresponding annotation files have been shared on the Figshare database54. Technical Validation The quality and reliability of the ‘Sungold’ T2T genome assembly were rigorously evaluated using multiple complementary approaches, focusing on the critical dimensions of continuity, completeness, and accuracy. For genome continuity, raw sequencing reads from the primary technologies (Illumina short-reads, PacBio HiFi reads, and ONT ultra-long reads) were comprehensively mapped back to the final assembly. This alignment was performed using the Burrows-Wheeler Aligner (BWA v0.7.17)55 for Illumina data and minimap2 (v2.24)56 for long-read datasets. The analysis yielded exceptionally high mapping rates ( >99.56%) across all platforms (Table 9). The proportion of the assembly covered at a depth of ≥ 20 × exceeded 99.97%, confirming near-complete and uniform sequence representation (Table 9). We have also calculated the coverage depth of the Illumina data using mosdepth (v0.3.12)57 with a window size of 100 kb, revealing a relatively comprehensive and uniform coverage (Fig. 6). Furthermore, the Hi-C heatmap of the P. armeniaca genome showed a high degree of consistency among all chromosomes, reflecting the accuracy of sequencing, ordering, and orientation of contigs in the genome assembly (Fig. 2). The structural integrity was further validated by the precise identification and mapping of all eight centromeres within their respective chromosomes, coupled with the detection of telomeric repeats at 14/16 chromosomal termini, demonstrating high completeness at the chromosome level (Fig. 4). Additionally, the long terminal repeat (LTR) assembly index (LAI) of the current assembly was 15.99, significantly exceeding the established “reference-quality” threshold (LAI > 10)58 (Table 1). For genome completeness, a comprehensive Benchmarking Universal Single-copy Orthologues (BUSCO v5.7.1)59 analysis against the embryophyta_odb10 database (containing 1,614 genes) revealed an overall completeness of 98.8% (1,594 genes found), comprising 96.5% single-copy (1,558), 2.2% duplicated (36), and 0.7% fragmented (11) BUSCOs (Fig. 7). To assess the quality of the annotation specifically, a protein-level BUSCO assessment was also conducted, showing 98.3% completeness (1,587 proteins found), with 96.4% single-copy (1,556), 1.9% duplicated (31), and 1.1% fragmented (17) orthologs (Fig. 7). The consistently high scores from BUSCO analyses provide strong evidence for the exceptional completeness of the assembled genome sequence and the accuracy of the gene models. For genome accuracy, the base-level accuracy of the assembly was quantified using Merqury (v1.3)60 to calculate a consensus Quality Value (QV). The resulting QV of 59.75 corresponds to an estimated base error of ~1 in 940,000 bases, confirming the high base-calling accuracy inherent in the ‘Sungold’ T2T assembly (Table 1). Additionally, we assessed the QV values for telomeric and centromeric regions, with mean values of 38.28 and 57.82, respectively, both lower than the whole-genome QV of 59.75. This aligns with the fact that centromeric and telomeric regions are highly repetitive and structurally complex, which poses inherent challenges for genome assembly and base-level accuracy. The observed lower QV values in these regions are therefore expected and consistent with the known difficulty of assembling repetitive sequences. Collectively, these multi-faceted evaluations—demonstrating outstanding continuity, completeness, and accuracy—firmly establish the high quality and reliability of the ‘Sungold’ T2T genome assembly as a foundational reference resource. Table 9Mapping rate and coverage statistics for different sequencing data types.Data setMean sequencing depthReads mappedBases mapped≥1×≥5×≥10×≥20×short reads116×99.83%98.83%100.00%100.00%99.99%99.97%ONT reads139×99.56%97.93%100.00%100.00%100.00%100.00%HiFi reads493×100.00%100.00%100.00%100.00%99.99%99.99% Mapping rate and coverage statistics for different sequencing data types. Coverage depth distribution of the Illumina data across the eight chromosomes of the Prunus armeniaca L. genome. The y-axis represents coverage depth; values exceeding the upper limit are truncated in the plot. The x-axis indicates genomic position in megabases (Mb). BUSCO assessments of assembly and annotation for the Prunus armeniaca L. genome. Acknowledgements This study was supported by the Agricultural Science and Technology Innovation Program (CAAS-ASTIP-ZFRI-05-202502), the Scientific Institution Basal Research Fund of ZFRI-CAAS (ZGS202309), the Strategic Co-operation Project between Chongqing Municipal People’s Government and Chinese Academy of Agricultural Sciences (CQ-CAAS-08), the National Natural Science Foundation of China (32172573), and the Forest Science and Technology Development Project of Henan Province (YLK202404). Author contributions K.C., P.W., R.W. and M.L. conceived the project. X.S., L.X., L.C. and Y.C. collected and prepared the plant materials. Z.H., H.B. and X.S. performed bioinformatics analysis. Z.H. wrote the original paper. K.C., P.W., R.W. and M.L. revised the original paper. All authors have read and approved the final manuscript. Data availability All data that support this study are accessible to the public. The sequencing datasets, which include Illumina reads, PacBio HiFi reads, ONT ultra-long reads, Hi-C data, and transcriptomic reads, have been deposited at the NGDC Genome Sequence Archive under the accession CRA02765650, and have also been deposited at the NCBI Sequence Read Archive under the accession SRP60284051. The T2T genome assembly has been deposited at the NGDC Genome Warehouse under the accession GWHIGUE00000000.152 and at the NCBI GenBank under the accession JBWBED00000000053. Additionally, the T2T genome assembly and its corresponding annotation files have been shared on the Figshare database54 (10.6084/m9.figshare.29465231.v1). Code availability All software and their specific versions used for data processing were clearly described in the Methods section. When specific parameters were not mentioned, the default settings were applied as guided. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Zhenyu Huang, Hangling Bie, Xinyue Su. Contributor Information Ruigang Wu, Email: wuruigang1986@126.com. Ming Li, Email: liming07@caas.cn. Pingyong Wang, Email: wangpingyong@caas.cn. Ke Cao, Email: caoke@caas.cn. References Associated Data Data Citations Data Availability Statement All data that support this study are accessible to the public. The sequencing datasets, which include Illumina reads, PacBio HiFi reads, ONT ultra-long reads, Hi-C data, and transcriptomic reads, have been deposited at the NGDC Genome Sequence Archive under the accession CRA02765650, and have also been deposited at the NCBI Sequence Read Archive under the accession SRP60284051. The T2T genome assembly has been deposited at the NGDC Genome Warehouse under the accession GWHIGUE00000000.152 and at the NCBI GenBank under the accession JBWBED00000000053. Additionally, the T2T genome assembly and its corresponding annotation files have been shared on the Figshare database54 (10.6084/m9.figshare.29465231.v1). All software and their specific versions used for data processing were clearly described in the Methods section. When specific parameters were not mentioned, the default settings were applied as guided.