A chromosome-level genome assembly of the native populations of the Japanese pavement ant, Tetramorium tsushimae Emery (Hymenoptera: Formicidae) Authors: Anandapadmanaban Gokulanathan, Sol-Moon Na, Soo-Jung Suh Published: 2026 ABSTRACT AbstractThe Japanes pavement ant, Tetramorium tsushimae Emery is an ecologically significant ant species recognized for its mutualistic interactions with aphids and other hemipterans. Here, we present the first chromosome-scale genome assembly for this species, generated using Oxford Nanopore long-read sequencing coupled with Pore-C scaffolding. The final assembly spans 259.2 Mb and is anchored to 13 chromosomes, with a scaffold N50 of 19.7 Mb. We identified 34,352 protein-coding genes, 58.9% of which were functionally annotated. The genome comprises 21.58% repetitive elements and exhibits 98.4% completeness based on BUSCO analysis. Gene family evolution analysis revealed orthogroup expansions and contractions, while synteny with Cardiocondyla obscurior Wheeler demonstrated extensive collinearity alongside lineage-specific rearrangements. This high-quality genome assembly provides a valuable foundation for future research into the evolution, ecology, and social behavior of T. tsushimae.Subject terms: Genome evolution, Genome assembly algorithms Background & Summary Tetramorium tsushimae Emery, commonly known as the Japanese pavement ant (JPA), is a social insect native to East Asia, particularly found in Japan and South Korea. It has also been recorded as an invasive species in North America1,2. Renowned for its ecological versatility, JPA thrives in both urban and natural environments due to its broad dietary habits and flexible nesting behavior3. JPA species plays vital ecological roles contributing to soil aeration, organic matter decomposition, and mutualistic interactions with aphids and other hemipterans, which collectively enhance nutrient cycling and biodiversity4,5. Its increasing presence in anthropogenic landscapes and its ability to exploit a wide range of ecological niches make JPA an emerging subject of interest for studies on ecological resilience, invasion biology, and social evolution. Despite its ecological importance, the lack of a high-quality reference genome for JPA has hindered functional genomic and evolutionary studies. While previous insect genomics studies have elucidated key mechanisms of host adaptation, environmental resilience, and pesticide resistance such as in Tetranychus piercei McGregor (Arachia: Tetrahychidae), Tuta absoluta Meyrick (Lepidoptera: Gelechiidae), and Dendrolimus punctatus Walker (Lepidoptera: Lasiocampidae) comprehensive genomic resources for socially complex ants like JPA remain limited6–8. This gap has constrained deeper understanding of the genetic mechanisms underlying ant adaptation, social behavior, and their interactions with host organisms and environments. Thus, a chromosome-level genome assembly for JPA is critically needed to unlock insights into its ecological dominance and potential applications in biodiversity and pest management. To address this gap, we generated a chromosome-scale genome assembly of JPA using a hybrid sequencing strategy. Oxford Nanopore long-read sequencing on the PromethION platform provided high-contiguity reads (12.98 Gb, N50 = 24.49 kb), while Pore-C sequencing was used for chromosome scaffolding via proximity ligation-based contact maps9. Genome size was estimated at 244.1 Mb using 21-mer analysis with Jellyfish10 and GenomeScope2.011, and the final assembled genome spanned 259.2 Mb, resolved into 13 chromosome-scale pseudomolecules with a scaffold N50 of 19.74 Mb. Assembly completeness was high, with 98.4% of BUSCO insecta genes recovered12. Gene prediction, integrating Illumina RNA-seq, and homology-based evidence from 20 ant species, identified 34,352 protein-coding genes. Functional annotation showed that 58.86% of predicted genes were assigned at least one functional term across NCBI nr (DIAMOND)13, InterProScan14, Gene ontology (GO)15, and KEGG databases16. Comparative analyses revealed strong macrosyntenic conservation with Cardiocondyla obscurior17, validating the structural accuracy of the assembled chromosomes. Orthogroup analysis and gene family evolution using OrthoFinder18 and CAFÉ19 further identified expansions in genes associated with detoxification, chemosensation, and thermal adaptation traits likely contributing to JPA’s ecological plasticity and success in urban environments. This chromosome-level genome represents the first high-resolution genomic resource for JPA, significantly advancing our capacity to explore the genetic basis of ant ecology, adaptation, and sociality. It provides a foundational platform for evolutionary genomics, synteny studies, and potential biocontrol applications, paving the way for broader investigations into the molecular ecology and evolution of social insects. Methods Sample collection, DNA extraction, and sequencing Whole-body pooled specimens of Tetramorium tsushimae were collected from field populations in the Republic of Korea (36°07′29.5″N 128 ° 11′57.1″E) during August 2024 and flash-frozen in liquid nitrogen. All sequencing libraries, including ONT genomic DNA, Pore-C, and RNA-seq datasets, were derived from pooled individuals collected from the same colony to maintain biological consistency across analyses. High-molecular-weight genomic DNA was extracted using a modified CTAB-based lysis method20. DNA was purified, and its integrity was verified via 1% agarose gel electrophoresis (Supplementary Figure 1). DNA concentration was measured using a Qubit 4 Fluorometer (Thermo Fisher Scientific,USA) with the Qubit dsDNA BR Assay Kit. The final yield was 12.9 µg with a concentration of 215 ng/µL. A Nanopore sequencing library was prepared using the Ligation Sequencing Kit SQK-LSK114 (Oxford Nanopore Technologies, Oxford, UK). Sequencing was performed on the PromethION platform using single FLO-PRO114M flow cell. Raw POD5 data were basecalled using Dorado v0.7.4.14 (https://github.com/nanoporetech/dorado), and reads with Phred quality scores ≥ 10 were retained. Quality metrics were assessed using NanoPlot21 v1.42.0 (Supplementary Figure 2). The read length distribution showed that most reads were below 50 kb, with ultra-long reads exceeding 100 kb. The average read quality scores were predominantly distributed between Q10 and Q20, with a peak around Q15–18, indicating high-quality long-read sequencing suitable for de novo genome assembly. Approximately 12.99 Gb of clean long-read data were obtained and used for genome assembly. RNA Sequencing for transcriptome profiling Transcriptome sequencing of Tetramorium tsushimae was conducted to aid in gene model prediction and functional annotation. Total RNA was extracted from pooled whole-body samples using the Qiagen RNeasy Mini Kit (Qiagen, Hilden, Germany). RNA integrity was assessed using an Agilent TapeStation, yielding a RIN score of 7.7. A stranded mRNA library was prepared using the TruSeq Stranded mRNA Library Kit (Illumina, San Diego, CA, USA) and sequenced on the Illumina NovaSeq. 6000 platform, generating 151 bp paired-end reads. Sequencing produced 78,470,386 read pairs, totaling 11.85 Gb of raw data (Table 1). Post-processing with Trimmomatic v0.39 retained high-quality reads, with 98.3% and 95.1% of bases exceeding Q20 and Q30, respectively. The average insert size was 359 bp, and the final library concentration was 14.9 ng/µL (55.4 nM). Table 1General information on sequencing raw data of this study.Read Type (Platform)Molecule TypeRaw Data Size (Gb)Cleaned Data Size (Gb)UseNCBI SRA AccessionNanopore long reads (PromethION)DNA12.9812.98De novo assemblySRR34053072Pore-C reads (PromethION)DNA16.1316.13Chromosome scaffoldingSRR34053073Illumina RNA-seq reads (NovaSeq. 6000)RNA11.811.8Gene predictionSRR34053074 General information on sequencing raw data of this study. Pore-C Sequencing for chromosomal contact information To generate chromatin conformation data for chromosome-level genome scaffolding, Pore-C libraries were prepared following established protocols22 and sequenced on the Oxford Nanopore Technologies (ONT) PromethION platform. Approximately 100 mg of whole-body tissue from Tetramorium tsushimae was flash-frozen, ground in liquid nitrogen, and cross-linked with 1% formaldehyde for 10 minutes at room temperature. Chromatin was permeabilized and denatured using 1% SDS at 65 °C, followed by digestion with NlaIII (New England Biolabs) for 18 hours at 37 °C. Proximity ligation was carried out using T4 DNA ligase at 16 °C for 6 hours to form chromatin concatemers. Cross-links were reversed by proteinase K treatment at 56 °C for 18 hours, and DNA was purified via phenol:chloroform extraction and ethanol precipitation. Sequencing libraries were constructed using the ONT Ligation Sequencing Kit SQK-LSK114, and sequencing was conducted on single FLO-PRO114M flow cell. Initial basecalling was performed in real time during sequencing using MinKNOW v23.11.5 for run monitoring and preliminary quality assessment. For downstream analyses, the raw POD5 files were subsequently re-basecalled offline using Dorado v0.7.4.14 to ensure consistent processing, standardized adapter trimming and quality filtering (Q ≥ 10), and to generate the final dataset used for genome assembly and scaffolding. Read quality metrics were assessed using NanoPlot21 v1.42.0, which revealed that the clean Pore-C dataset totaled 16.13 Gb, comprising 24.4 million reads, with a mean read length of 660 bp and an N50 of 905 bp. As expected for Pore-C libraries, a substantial proportion of bases (45%) were contributed by reads longer than 1 kb due to the concatemeric nature of proximity-ligated fragments. De Novo genome assembly and scaffolding with pore-C reads To reconstruct the chromosome-scale genome of JPA, we combining long-read sequencing with chromatin conformation scaffolding. High-coverage ONT long-read sequencing generated 12.98 Gb of filtered genomic data, with a read N50 of 24,492 bp. Basecalling and quality filtering were performed as described above. These long reads were used for de novo assembly with NextDenovo v2.5.223, producing an initial assembly of 57 contigs, totaling 259.2 Mb, with a contig N50 of 10.39 Mb and a maximum contig length of 20.5 Mb. Assembly completeness was assessed using BUSCO v5.5.012 with the insecta_odb10 dataset (n = 1,367), which indicated 98.8% completeness. To provide a more lineage-specific evaluation, BUSCO analysis was additionally performed using the hymenoptera_odb10 dataset (n = 5,991), yielding 99.5% completeness at the genome level. Genome size was estimated using 21-mer frequency analysis with Jellyfish 2.2.310, followed by modeling in GenomeScope 2.024 under diploid assumptions (Fig. 1A). The estimated haploid genome size was 244.1 Mb, with a heterozygosity rate of 0.43%, unique content of 81.1%, and a sequencing error rate of 0.23%. The model fit (94.39%) and k-mer coverage peak (31 × ) were consistent with the observed long-read assembly statistics. For chromosomal scaffolding, the Pore-C data described above (16.13 Gb) were processed using the Pore-C-Snakemake pipeline v0.4.025, which converts multi-contact reads into pairwise interaction maps. The resulting multi-contact data were converted into pairwise contact maps and input into the 3D-DNA pipeline v19071626 using the following parameters: -i 50000–polisher-input-size 1000000–splitter-input-size 1000000 -r 2–editor-coarse-resolution 25000–editor-coarse-region 125000–polisher-coarse-resolution 100000–polisher-coarse-region 3000000. The draft scaffolds were further curated using Juicebox v1.11.0827 a Hi-C contact visualization tool, to identify and resolve misassemblies. The resulting Pore-C contact map showed strong intra-chromosomal interaction signals along the diagonal, confirming the correct ordering and orientation of the chromosomal scaffolds (Supplementary Figure 3). The final output consisted of 13 chromosome-scaffold with 98.3% of the genome anchored and a scaffold N50 improved to 19.74 Mb, achieving a chromosome-level assembly suitable for downstream functional and comparative genomic analyses. The assembly results are presented in Table 2. The chromosomal genome characteristics and self-collinearities were profiled using reciprocal BLAST searches, and the results were visualized using Circos v0.6928 (Fig. 1C). Chromosome-scale de novo genome assembly of T. tsushimae. (A) Genome size estimation based on 21-mer frequency analysis. (B) Circos plot of the assembled JPA genome. From innermost to outermost, the concentric tracks represent: GC% - GC percentage, Repeat(%)- repeat element density, Gene(#) - gene numbers, tRNA, rRNA and interspecific co-linearity of protein-coding genes relationship in red color. Table 2Genome assembly and scaffolding statistics.Assembly StatisticsInitial AssemblyAfter Pore-C ScaffoldingFinal Chromosome-Scale AssemblyAssembly size (bp)259,200,000 (ONT contigs)259,200,000 (scaffolded)254,800,000 (13 Chromosome)Number of contigs/scaffolds (>10 kb)57 contigs57 scaffolds13 scaffolds + 14 unscaffoldedN50 length (bp)10,390,00019,740,00019,740,000Longest contig/scaffold (bp)20,500,00024,900,00024,900,000GC content (%)37.4737.4737.47 Genome assembly and scaffolding statistics. Gene prediction and functional annotation Prior to gene prediction, repetitive elements in the JPA genome were identified and masked using RepeatModeler v2.0.529, LTR_Finder_parallel v1.230, LTRharvest v1.6.531, LTR_retriever v2.9.932, and RepeatMasker v4.1.633, employing default parameters and soft-masking. Repeat annotation revealed that 21.58% of the assembled genome consists of repetitive sequences. Gene prediction was conducted using the BRAKER3 v3.0.7 pipeline34,35 integrating ab initio prediction with transcriptomic and protein homology evidence. Within this framework, GeneMark-ETP36 was used for the initial gene model predictions, followed by training with AUGUSTUS37 and integration via TSEBRA38. BUSCO v5.5.012 (lineage: insecta) was used to assess annotation completeness, revealing 34,352 predicted protein-coding genes with a BUSCO completeness score of 98.5%, including 97.7% single-copy genes. The total predicted gene length was approximately 30.4 Mb. Transcriptomic evidence from Illumina RNA-seq was incorporated to refine gene model structures. Illumina reads were quality-filtered using Trimmomatic v0.3939 and aligned to the soft-masked genome using HISAT2 v2.2.140 to identify splice junctions. While a de novo transcriptome assembly using Trinity v2.1541 was also performed, the integration of transcript evidence was handled directly within the BRAKER3 v3.0.7 framework, without the use of PASA or other isoform refinement tools. Protein homology evidence was integrated into BRAKER3 v3.0.7 using annotated protein sequences from 20 ant species retrieved from the NCBI Genome Database and relevant publications. These reference species included Solenopsis invicta, Monomorium pharaonis, Cardiocondyla obscurior, Temnothorax longispinosus, Acromyrmex echinatior, Wasmannia auropunctata, Pogonomyrmex barbatus, Trachymyrmex septentrionalis, Temnothorax nylanderi, Cyphomyrmex costatus, Trachymyrmex zeteki, Acromyrmex insinuator, Temnothorax curvispinosus, Vollenhovia emeryi, Trachymyrmex cornetzi, Atta colombica, Atta cephalotes, Acromyrmex charruanus, Pseudoatta argentina, and Acromyrmex heyeri. These proteomes enabled the identification and refinement of conserved gene models (Fig. 2B). The accession numbers and related details of reference species used for protein homology analysis are provided in the Supplementary Table 1. Genome annotation summary. Distribution of annotated genes across functional categories, highlighting the proportion of curated, experimental, hypothetical, and uncharacterized proteins (A). Top matching species based on nr BLAST hits, providing insight into the phylogenetic origin and homology of annotated proteins (B). Abundance of conserved protein domains identified by InterPro, representing functional and structural motifs across the gene set (C). Classification of annotated genes into major GO categories: Biological Process (BP), Molecular Function (MF), and Cellular Component (CC) (D). Distribution of annotated genes across KEGG functional categories, color-coded by high-level pathway class (E). Functional annotation was performed using a combination of databases and tools. DIAMOND v2.1.813 was used to align predicted proteins against the NCBI non-redundant (nr) database, resulting in 20,183 annotated genes (58.75%) (Fig. 2A). InterProScan v5.64 provided protein domain annotations for 12,437 genes (36.20%)14 (Fig. 2C). GO terms were assigned using BLAST2GO15, annotating 12,832 genes (37.35%) (Fig. 2D). KEGG pathway mapping was conducted using the KEGG Automatic Annotation Server (KAAS) identifying 6,831 genes (19.89%) with associated pathway information16 (Fig. 2E). In total, 20,219 protein-coding genes (58.86% of the full set) were functionally annotated based on at least one database, supporting downstream functional and evolutionary analyses(Table 3). Table 3Detailed information for the T. tsushimae gene model.Gene model statisticsValueNumber of predicted genes34,352Number of predicted proteins34,645Complete BUSCOs (Insecta, genome mode)96.5% (S: 62.0%, D: 34.5%)Complete BUSCOs (Insecta, protein mode)92.4% (S: 70.1%, D: 22.3%)Number of database-hit genesInterPro12,437NCBI nr (DIAMOND)20,183EggNOG12,832KEGG (KAAS)6831 Detailed information for the T. tsushimae gene model. Orthologous gene profiling and synteny analysis Orthologous gene relationships and genome evolution were analyzed across JPA and other Hymenoptera species (Fig. 3A). Orthogroups were inferred using OrthoFinder v2.5.4 with default parameters18. For phylogenetic analysis, 300 single-copy orthologs identified by OrthoFinder were aligned using MAFFT v7.52642, and a species tree was reconstructed using BEAST v2.6.743. Single BEAST runs were conducted under a Gamma Site Model (4 categories, JTT substitution model) with a strict molecular clock and an MCMC chain length of 10,000,000 generations each. Convergence and effective sample size (ESS) values were assessed using Tracer (ensuring all parameters achieved an ESS > 200), and a 10% burn-in was applied prior to tree summarization. To obtain a time-calibrated ultrametric tree for CAFE analysis, divergence time estimates were derived from the TimeTree database (www.timetree.org) for relevant Hymenoptera lineages, and the corresponding node was calibrated using the reported median divergence time. The calibrated tree was subsequently used as input for CAFÉ v5.1.0 to assess gene family expansions and contractions under default settings44. Syntenic relationships between JPA and the reference genome Cardiocondyla obscurior were explored using MCScanX v1.0.045, identifying conserved collinear blocks across chromosomes (Fig. 3B). Visualization of genome-wide syntenic relationships and chromosomal rearrangements was carried out using SynVisio46. The Orthologous gene profile of T. tsushimae. Orthologous gene gain and loss across the T. tsushimae genome and other reference insects. Numbers at each internal node or terminal branch represent the number of gene families gained (green) and lost (red) relative to the ancestral node (A). Synteny alignment between the chromosomes of T. tsushimae (Tt01–Tt13) and the reference genome of Cardiocondyla obscurior (Co01–Co29), illustrating conserved chromosomal collinearity and rearrangements (B). Data Records The raw sequencing data have been deposited in the NCBI Sequence Read Archive (SRA) under study accession SRP59303847. Nanopore long-read sequencing, Pore-C sequencing, and transcriptome sequencing raw reads were deposited in the SRA database with accession numbers SRR3405307248, SRR3405307349, and SRR3405307450. The chromosomal genome assembly of T. tsushimae has been deposited in NCBI GenBank under the accession number JBXGQE000000000.151. Technical Validation Quality control of nucleic acids and sequencing libraries The quality and purity of extracted DNA were assessed by 1% agarose gel electrophoresis and Nanodrop spectrophotometry (A260/A280 and A260/A230), while DNA quantity was measured using the Qubit in conjunction with the dsDNA High Sensitivity Assay Kit. RNA integrity was evaluated using the TapeStation RNA Screen Tape (Agilent Technologies, USA). Additionally, for Pore-C sequencing, the quantity of proximity-ligated DNA was confirmed to be 12.9 μg, with a concentration of 215 ng/μL. All Illumina sequencing libraries underwent quality checks using the TapeStation D1000 Screen Tape (Agilent Technologies, USA), ensuring the suitability of nucleic acid libraries for downstream sequencing applications. Genome assembly and annotation validation The contiguity and completeness of the Tetramorium tsushimae genome assembly were evaluated using multiple metrics. The final ONT-based de novo assembly consisted of 57 contigs spanning 259.2 Mb with a contig N50 of 10.39 Mb. Scaffolding with Pore-C data using 3D-DNA v190716 increased the scaffold N50 to 19.74 Mb, anchoring 98.3% of the genome into 13 chromosome-scale pseudomolecules. Nanopore PromethION reads (12.99 Gb) were aligned to the final assembly using Minimap2 v2.26 for structural validation and coverage inspection. The alignment confirmed even coverage across the genome, supporting assembly reliability. Assembly completeness was assessed using BUSCO v5.5.0 with the insecta_odb10 lineage dataset in both genome (-m genome) and protein (-m proteins) modes, which revealed completeness scores of 98.4% and 98.5%, respectively. Gene prediction was performed using BRAKER v3.0.7, integrating RNA-seq data (Illumina paired-end, 151 bp, 78.5 million reads) aligned with HISAT2 v2.2.1. The final gene set included 34,352 protein-coding genes. Functional annotation was conducted with DIAMOND v2.1.9, InterProScan v5.66, and BLAST2GO v5.2.5, confirming gene model accuracy and completeness. Supplementary information Supplementary figures Supplementary Table 1. Acknowledgements This work was supported by the Ministry of agriculture, Food and Rural Affairs Quarantine and Inspection Technology Development Project (Grant number:I-1543086-2024-26-01). Author contributions S.J.S. conceived the project. S.J.S. and A.G. performed the field collection. A.G. wrote the original manuscript. S.J.S. and S.M.N. revised the manuscript. A.G. performed genome assembly, annotation, and figure generation using R programming. Data availability The datasets generated in this study are publicly available. Raw sequencing data have been deposited in the NCBI Sequence Read Archive under study accession SRP593038, including Nanopore long-read sequencing (SRR34053072), Pore-C sequencing (SRR34053073), and transcriptome sequencing reads (SRR34053074). The chromosome-level genome assembly and genome annotation of T. tsushimae are available in the NCBI datasets and the Figshare repository. Code availability All software and pipelines used in this genome study were executed in accordance with the developers’ manuals, as provided in the respective bioinformatics publications or official sources such as GitHub. Software versions and detailed parameters are listed in the Methods section; default settings were applied unless otherwise specified. For additional scripts or code, please contact A.G. 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. Supplementary information The online version contains supplementary material available at 10.1038/s41597-026-07426-6. References Associated Data Data Citations Supplementary Materials Supplementary figures Supplementary Table 1. Data Availability Statement The datasets generated in this study are publicly available. Raw sequencing data have been deposited in the NCBI Sequence Read Archive under study accession SRP593038, including Nanopore long-read sequencing (SRR34053072), Pore-C sequencing (SRR34053073), and transcriptome sequencing reads (SRR34053074). The chromosome-level genome assembly and genome annotation of T. tsushimae are available in the NCBI datasets and the Figshare repository. All software and pipelines used in this genome study were executed in accordance with the developers’ manuals, as provided in the respective bioinformatics publications or official sources such as GitHub. Software versions and detailed parameters are listed in the Methods section; default settings were applied unless otherwise specified. For additional scripts or code, please contact A.G.