Chromosomal-level genome assembly of Trypanosoma carassii, the etiologic agent of a recent outbreak of trypanosomiasis in cage-cultured large yellow croaker (Larimichthys crocea) in China Authors: Nai-Cheng Liu, Xiao-Rong Ji, Jin-Yong Zhang, Zhao-Zhe Xin Published: 2026 ABSTRACT Trypanosoma carassii, a typical freshwater fish trypanosome, has recently been identified as the etiological agent of a trypanosomiasis outbreak in cage-cultured large yellow croaker (Larimichthys crocea) in China and has been designated as T. c. larimichthys. To date, publicly available genomic data for trypanosomes have been limited to terrestrial species, particularly those of medical importance. Here, we present a chromosome-level genome assembly of T. carassii, the first genome of an aquatic trypanosome, generated using PacBio HiFi long-read sequencing and Hi-C scaffolding technologies. A preliminary genome survey based on Illumina sequencing data estimated the genome size at 56.38 Mb with a heterozygosity of 1.17%. The final assembled genome spans 48.55 Mb, with contig N50 and scaffold N50 values of 139.15 Kb, and achieves 100.00% BUSCO completeness. Hi-C data resolved the assembly into 34 chromosomes and 9 unanchored scaffolds. Repetitive elements account for 53.29% of the genome (approximately 25.87 Mb). A total of 11,584 protein-coding genes were predicted, 95.36% of which were functionally annotated. Synonymous substitution rates analysis of paralogous genes indicates a recent burst of gene duplication, which likely corresponds to a whole-genome duplications. This high-quality genome assembly provides invaluable resources for understanding the evolution and host adaptation of aquatic trypanosomes. Background & Summary Trypanosomes are unicellular parasitic haemoflagellates that infect a broad range of vertebrate hosts, including fish, reptiles, amphibians, birds, and mammals. These parasites are well-known pathogens responsible for devastating diseases such as fish trypanosomiasis, Surra and Nagana in domestic and wild mammals, as well as African sleeping sickness and Chagas disease in humans1,2. In aquaculture, trypanosomiasis has been frequently documented in various freshwater and marine fish species, including cultured groupers (Cromileptes altivelis and Epinephelus fuscoguttatus)3, largemouth bass (Micropterus salmoides)4, Amazon sailfin catfish (Pterygoplichthys pardalis)5, Nile tilapia (Oreochromis niloticus)6, blood parrot cichlid (Vieja melanura♀ × Amphilophus citrinellus♂)7. A recent severe outbreak of trypanosomiasis has occurred in farmed large yellow croaker (Larimichthys crocea) in Southeast China since 2023, resulting in catastrophic economic losses8–10. The annual cumulative mortality caused by this emerging trypanosomiasis in Fujian province, the main production area of big yellow croaker in China was above 50% in 2023 and 2024. To date, no effective prophylactic or therapeutic approaches are available. Trypanosoma carassii was initially discovered and described in the blood of the freshwater crucian carp (Carassius carassius)11. It was subsequently identified as the causative agent of trypanosomiasis in both cyprinid and non-cyprinid freshwater fish species12. The broad host range of T. carassii encompasses multiple Orders of freshwater fish, including Cypriniformes, Perciformes, and Siluriformes13. The identification of T. carassii in the L. crocea represents the first documented case of this freshwater trypanosome infecting a marine fish host8–10. This indicates that T. carassii can breach the barriers between the internal environments of marine and freshwater fish, thereby enabling survival in diverse habitats. The host adaptation strategies of trypanosomes are inherently complex and intimately linked to their genomic architecture. Substantial advances have been achieved in the genome sequencing and analysis of medically significant mammalian trypanosomes, including T. brucei14, T. rangeli15, T. conorhini15, T. grayi16, T. b. gambiense17, T. cruzi18–20, T. evansi21, T. theileri22, T. equiperdum23, and T. melophagium24. In contrast, the genomic landscape of fish-infecting trypanosomes remains largely unexplored and poorly characterized. This knowledge gap seriously hinders our understanding of the basic biological processes of fish parasites, including species classification, genetic diversity, host specificity and adaptability. Here, we present the first high-quality chromosome-level genome assembly of aquatic trypanosome, generated using a combination of PacBio HiFi long-read sequencing and Hi-C scaffolding technologies. A preliminary genome survey based on Illumina sequencing data estimated the genome size at 56.38 Mb with a heterozygosity of 1.17%. The final T. carassii genome assembly comprises 34 chromosomes and 9 unanchored scaffolds. The total genome size is 48.55 Mb, with a GC content of 45.33%. The contig N50 and scaffold N50 values are both 139.15 Kbp, and the chromosome anchoring rate reaches 98.08% (Table 1). Genomic component annotation mainly includes repetitive sequence prediction, gene structure prediction, and non-coding RNA prediction. Functional annotation is to infer the function of genes through homology alignment. A total of 11,584 genes were identified in the T. carassii genome, of which 11,046 (95.36%) were functionally annotated using public databases including NR, Swiss-Prot, KEGG, GO, and eggNOG (Table 2). High-quality genomic data provide valuable resources for understanding the evolution and host adaptation of aquatic trypanosomes.Table 1Summary of the T. carassii genome assembly.Sample nameScaffold numberScaffold length (bp)GC (%)Contig N50 (bp)ScaffoldN50 (bp)Chromosome percent (%)Trypanosoma carassii4348,548,46945.331,391,4921,391,49298.08Table 2Summary statistics of gene structure prediction and function annotation.StatisticsNumber&LengthGene structure prediction Number of genes11,584 Total genic length20,466,425 Mean gene length1,766 Number of transcripts11,584 Transcripts per gene1 Total transcript length20,466,403 Mean transcript length1,766 Number of exons11,585 Exons per transcript1 Mean exon length1,766 Number of coding exons11,585 Number of introns1 Mean intron length22 Total cds length19,033,979 Mean CDS length1,643Gene function annotation Protein-coding genes11,584 Annotated genes11,046 Annotated by NR11,033 Annotated by SwissProt4,852 Annotated by KEGG3,642 Annotated by GO785 Annotated by eggNOG7,423 Not Annotated538 Table 1Summary of the T. carassii genome assembly.Sample nameScaffold numberScaffold length (bp)GC (%)Contig N50 (bp)ScaffoldN50 (bp)Chromosome percent (%)Trypanosoma carassii4348,548,46945.331,391,4921,391,49298.08 Summary of the T. carassii genome assembly. Table 2Summary statistics of gene structure prediction and function annotation.StatisticsNumber&LengthGene structure prediction Number of genes11,584 Total genic length20,466,425 Mean gene length1,766 Number of transcripts11,584 Transcripts per gene1 Total transcript length20,466,403 Mean transcript length1,766 Number of exons11,585 Exons per transcript1 Mean exon length1,766 Number of coding exons11,585 Number of introns1 Mean intron length22 Total cds length19,033,979 Mean CDS length1,643Gene function annotation Protein-coding genes11,584 Annotated genes11,046 Annotated by NR11,033 Annotated by SwissProt4,852 Annotated by KEGG3,642 Annotated by GO785 Annotated by eggNOG7,423 Not Annotated538 Summary statistics of gene structure prediction and function annotation. Methods Genomic DNA extraction, library construction, and sequencing Genomic DNA was extracted using the Blood & Cell Culture DNA Kit (Qiagen, 13323) according to the manufacturer’s instructions. Briefly, samples were lysed and proteins were denatured in a lysis buffer. After adding protease and incubating, the lysates were loaded onto a QIAGEN Genomic-tip. The DNA bound to the column while other cellular components were washed away. After a wash step to remove contaminants, pure, high-molecular-weight DNA was eluted. Finally, DNA quality was assessed using a NanoDrop (Thermo Fisher Scientific, ND-2000), a Qubit fluorometer (Invitrogen, Q33238), and electrophoresis. After examination of the quality of isolated DNA, the library of 15 Kb was constructed using a SMRTbell prep kit 3.0 (Pacific Biosciences, 102-182-700). The construction includes DNA shearing, DNA damage reparation, end-repair, A-tailing, SMRTbell adapter ligation, nuclease treatment, and size selection of the library. After quality control test by Agilent Fragment Analyzer System (Agilent Technologies, M5310AA), a SMRTbell library was obtained. The library was sequenced on the PacBio Revio system (Pacific Biosciences, 102-090-600). Illumina sequencing and genome size estimation RNA samples from two developmental stages (high-density and low-density growth phase) were extracted using TRIzol reagent (Invitrogen, USA). RNA-seq libraries were constructed using the Illumina TruSeq Stranded mRNA Library Prep Kit (Illumina, USA) and sequenced by the Illumina NovaSeq. 6000 platform. The clean data was obtained by removing reads containing adapters, reads containing poly-N, and low-quality reads from the raw data. HISAT2 v2.2.125 (-q --dta-cufflinks -x) was used to align clean data to the genome sequence. The genome size of T. carassii was estimated by the k-mer method26 using sequencing data from illumina DNA library. K-mer distribution was estimated by using jellyfish v1.1.1127 with default parameters. The heterozygosity ratio was estimated by the online tool of GenomeScope v128 with default parameters (https://github.com/schatzlab/genomescope). Finally, the genome size was calculated according to the formula that genome size = K-mer coverage/Mean k-mer depth and the result showed that genome size is 56.38 Mb, heterozygosity is 1.17% (Fig. 1A).Fig. 1Sequencing, survey and assembly of T. carassii genome. (A) K-mer distribution graph, the horizontal axis represents the Kmer coverage depth, while the vertical axis shows the frequency of Kmers corresponding to that depth. (B) the distribution of HiFi reads number with length. (C) The distribution of base number with length of HiFi reads. (D) Hi-C contact map of the T. carassii genome. The blocks represent the contacts between one location and another. The color illustrates the contact density from red (high) to low (orange). (E) Intragenomic collinearity analysis, the outermost ring represents 34 chromosomes and 9 scaffolds, the inner circle represents the GC content of 34 chromosomes and 9 scaffolds, and the innermost circle is represented by different-sized squares, which correspond to the sizes of chromosomes and scaffolds. Sequencing, survey and assembly of T. carassii genome. (A) K-mer distribution graph, the horizontal axis represents the Kmer coverage depth, while the vertical axis shows the frequency of Kmers corresponding to that depth. (B) the distribution of HiFi reads number with length. (C) The distribution of base number with length of HiFi reads. (D) Hi-C contact map of the T. carassii genome. The blocks represent the contacts between one location and another. The color illustrates the contact density from red (high) to low (orange). (E) Intragenomic collinearity analysis, the outermost ring represents 34 chromosomes and 9 scaffolds, the inner circle represents the GC content of 34 chromosomes and 9 scaffolds, and the innermost circle is represented by different-sized squares, which correspond to the sizes of chromosomes and scaffolds. De novo and Hi-C-assisted genome assembly The PacBio SMRT-Analysis package was used for the quality control of the raw reads. We removed the following types of reads: (i) reads with the quality value lower than 0.8; (ii) reads containing the adaptor ligation to itself; and removed the adaptor sequence in the reads. HiFi reads were generated by SMRTLink software with parameters–min-passes = 3–min-rq = 0.99. A total of 108 Gb (~193 × ) PacBio HiFi reads were generated, the length of reads ranges from 272 bp to 55,472 bp (Table 3). The length distribution of HiFi reads is illustrated in Fig. 1B, while the corresponding base distribution is depicted in Fig. 1C. Hifiasm v0.19.529 with default parameters was applied to generate the draft assembly with HiFi reads. De novo assembly yielded 107 primary contigs (Table 4). Purge_dups v1.2.530 was used to remove heterozygous sequences, and the parameters used are −2 -T changed.cutoffs -l XXX –m XXX -u XXX. The resulting genomic assembly obtained through this step is presented in Table 4. We cut the contigs with assembled genomes smaller than 10 Mb and compared them with the NT database to determine if there was any contamination from other species. The results showed no contamination, so this step did not filter out the contigs (Table 4). We removed chimeric contigs with an average coverage depth of less than 30 or with an average GC content outside the range of 0.00% to 60.00% (Table 4).Table 3Statistics of HiFi reads results.Sample NameReads NumberBase NumberMin LengthMean LengthMax LengthN50 (bp)Depth (X)Trypanosoma carassii744,69010,839,546,03827214,55555,47214,523193Table 4Sequential optimization statistics of the T. carassii genome assembly.Sample NameContig NumContig Len(bp)GC (%)Contig Min (bp)Contig N50 (bp)Contig N90 (bp)Contig Max (bp)Hifiasm genome assemblyTrypanosoma carassii10761,815,76545.051,8921,440,386935,3494,138,069Statistics of genome assembly results after heterozygosity purgingTrypanosoma carassii7360,686,73445.051,8921,440,386940,0614,138,069Statistics of genome assembly results after decontaminationTrypanosoma carassii7360,686,73445.051,8921,440,386940,0614,138,069Statistics of genome assembly results after chimera removalTrypanosoma carassii5160,083,17845.0346,2271,440,386966,7644,138,069 Table 3Statistics of HiFi reads results.Sample NameReads NumberBase NumberMin LengthMean LengthMax LengthN50 (bp)Depth (X)Trypanosoma carassii744,69010,839,546,03827214,55555,47214,523193 Statistics of HiFi reads results. Table 4Sequential optimization statistics of the T. carassii genome assembly.Sample NameContig NumContig Len(bp)GC (%)Contig Min (bp)Contig N50 (bp)Contig N90 (bp)Contig Max (bp)Hifiasm genome assemblyTrypanosoma carassii10761,815,76545.051,8921,440,386935,3494,138,069Statistics of genome assembly results after heterozygosity purgingTrypanosoma carassii7360,686,73445.051,8921,440,386940,0614,138,069Statistics of genome assembly results after decontaminationTrypanosoma carassii7360,686,73445.051,8921,440,386940,0614,138,069Statistics of genome assembly results after chimera removalTrypanosoma carassii5160,083,17845.0346,2271,440,386966,7644,138,069 Sequential optimization statistics of the T. carassii genome assembly. The Hi-C library was constructed through the following steps: crosslinking cells with formaldehyde, digesting DNA with MboI, filling ends and mark with biotin, ligating the resulting blunt-end fragments, purification and random shearing DNA into 300–500 bp fragments. After quality control test of the libraries using Qubit 2.0, an Agilent 2100 instrument (Agilent Technologies, CA, USA) and QPCR, 150 bp PE sequencing of the Hi-C library were performed on the Illumina Novaseq. 6000 platform. A total of 27,499,406 pairs of sequencing reads were obtained, amounting to 8,249,821,800 bp (approximately 8.25 Gb) of clean bases (Table 5). To minimize artificial bias, we removed the following types of reads: (i) reads with 3nt unidentified nucleotides (N); (ii) reads aligned to the adaptor; (iii) reads with ≥ 20% bases having phred quality ≤ 5. The filtered Hi-C reads were aligned to the initial draft genome by BWA v0.7.1731 with default parameters which was integrated into Juicer v1.6.232. Only uniquely mapped and valid paired-end reads were used to assembly by 3D-DNA 18092233. Juicebox v1.9.834 was used to manually order the scaffolds to get the final chromosome assembly. Contact maps were plotted with HiCExplorer v3.335. From a statistical perspective, the sequencing error rate is low, indicating high data quality. The GC content is consistent with the genomic characteristics of this species, and no N bases were detected, suggesting thorough data filtering with no ambiguous base residues. The number of read pairs effectively aligned to the genome was 19,831,389, with an alignment success rate of 72.12%, indicating that the majority of the sequencing data can be effectively used for subsequent analysis (Table 5).Table 5Hi-C data quality control and comparison results.Sample NameClean reads (pair)Clean bases (bp)Clean N (%)Clean GC (%)Clean Q20 (%)Clean Q30 (%)Alignable pairedTrypanosoma carassii27,499,4068,249,821,8000.00;0.0044.59;44.9098.87;98.1496.47;94.6319,831,389 (72.12%) Table 5Hi-C data quality control and comparison results.Sample NameClean reads (pair)Clean bases (bp)Clean N (%)Clean GC (%)Clean Q20 (%)Clean Q30 (%)Alignable pairedTrypanosoma carassii27,499,4068,249,821,8000.00;0.0044.59;44.9098.87;98.1496.47;94.6319,831,389 (72.12%) Hi-C data quality control and comparison results. Due to the complexity of the species, distinguishing chromosome boundaries in the Hi-C analysis proved challenging, and no interaction between contigs was observed. During assembly, each contig initially represented a chromosome in the Hi-C map, resulting in identical contig N50 and scaffold N50 values. During the Hi-C stage, we found some heterozygous fragments that could not be removed by the purge_dups software. Their representation on the heatmap was similar to the diagonal lines of a polyploid heatmap. We manually removed these heterozygous fragments, thus the genome size, contig number, and contig N50 decreased. Based on the above reasons, the final T. carassii genome assembly comprises 34 chromosomes and 9 unanchored scaffolds (Fig. 1D,E). The total genome size is 48.55 Mb. The contig N50 and scaffold N50 values are both 139.15 Kbp, and the chromosome anchoring rate reaches 98.08% (Table 1). Assessment of the genome assembly quality To assess the accuracy of the genome assembly sequence, HiFi reads were aligned to the draft genome assembly with Minimap v2.26 (-I6G -ax map-pb -t 60)36. We employed the mosdepth v0.3.1 (-t 64)37 to calculate genome coverage. Re-alignment of the sequencing reads to the assembled genome demonstrated a mapping rate of 99.77%, and genome coverage is 100.00% (Table 6). Based on the alignment results, coverage depth and GC content were calculated in 10-kb windows using SAMtools v1.1738 and bedtools v2.2.2939 with default parameters. The GC-Depth plot was then generated using R. The GC content distribution and coverage depth across the genome are presented in Fig. 2A and Fig. 1B, respectively. Figure 2A shows a distinct bimodal GC distribution. Figure 2B does indeed indicate the presence of multiple DNA populations. We speculated that the T. carassii genome has different genomic compartments, possibly related to mitochondrial DNA (kDNA), including minicircles and maxicircles. These structures have significantly different GC content and coverage from nuclear DNA. Kinetoplast (k) DNA is a specific form of mitochondrial DNA in trypanosomes, consisting of minicircles and maxicircles. It is crucial for the survival of these parasitic protists40. The final GC content of the genome is 45.33%. Genome assembly and completeness were assessed using the conserved genes in the benchmarking universal single-copy orthologs (BUSCO) database, with the euglenozoa_odb10 dataset41. The analysis revealed that the genomic completeness was 100.00% (83.1% single-copy, 16.9% duplicated) (Table 7).Table 6HiFi data back ratio and genomic coverage statistics.Sample NameRead NumberMapped ReadsPercent(%)Average DepthCoverageTrypanosoma carassii744,690742,94699.77178.92100.00%Fig. 2Genome assessment, annotation and whole-genome duplication analysis of T. carassii. (A) The distribution of GC content, the horizontal axis represents the GC content (%) and the vertical axis represents the proportion of DNA fragments (or genomic windows) with a certain GC content among all the fragments. (B) GC content and depth distribution, the horizontal axis represents sequencing depth, and the vertical axis represents the percentage of GC content. (C) Venn diagram of gene function annotation in different databases, including NR, Swiss-Prot, GO, eggNOG, KEGG. (D) The eggNOG function classification of consensus sequence of the T. carassii genome. (E) The gene ontology (GO) analysis of the T. carassii genome. (F) Analysis of the whole-genome duplications (WGD) of T. carassii. Each peak shown in the figure represents a WGD event. The horizontal axis represents the distribution of Ks values of homologous gene (paralogous) pairs, and the vertical axis represents the number of homologous gene (paralogous) pairs. The peak (Ks≈0.0176) indicated a recent burst of gene duplication in the genome of T. carassii, which likely corresponds to a WGD.Table 7Assessment of genome assembly integrity.Sample NameBUSCOTrypanosoma carassiiC: 100.0% [S: 83.1%, D: 16.9%], F: 0.0%, M: 0.0%, n: 130 Table 6HiFi data back ratio and genomic coverage statistics.Sample NameRead NumberMapped ReadsPercent(%)Average DepthCoverageTrypanosoma carassii744,690742,94699.77178.92100.00% HiFi data back ratio and genomic coverage statistics. Genome assessment, annotation and whole-genome duplication analysis of T. carassii. (A) The distribution of GC content, the horizontal axis represents the GC content (%) and the vertical axis represents the proportion of DNA fragments (or genomic windows) with a certain GC content among all the fragments. (B) GC content and depth distribution, the horizontal axis represents sequencing depth, and the vertical axis represents the percentage of GC content. (C) Venn diagram of gene function annotation in different databases, including NR, Swiss-Prot, GO, eggNOG, KEGG. (D) The eggNOG function classification of consensus sequence of the T. carassii genome. (E) The gene ontology (GO) analysis of the T. carassii genome. (F) Analysis of the whole-genome duplications (WGD) of T. carassii. Each peak shown in the figure represents a WGD event. The horizontal axis represents the distribution of Ks values of homologous gene (paralogous) pairs, and the vertical axis represents the number of homologous gene (paralogous) pairs. The peak (Ks≈0.0176) indicated a recent burst of gene duplication in the genome of T. carassii, which likely corresponds to a WGD. Table 7Assessment of genome assembly integrity.Sample NameBUSCOTrypanosoma carassiiC: 100.0% [S: 83.1%, D: 16.9%], F: 0.0%, M: 0.0%, n: 130 Assessment of genome assembly integrity. Genome components and functional annotation Genomic component annotation mainly involves classifying sequences in the genome, including repetitive sequence prediction, gene structure prediction, and non-coding RNA prediction. Functional annotation is to infer the function of genes through homology alignment. The combination of de novo and homology-based method was used to predict repeat elements. For de novo annotation, MITE-Hunter v1.042 (-n 5 -c 20 -S 12345678) was used to identify the mini-inverted repeat transposable elements (MITEs) belonging to type II transposable elements. LTRharvest43 and LTR Finder v1.0744 (-D 15000 -d 1000 -L 7000 -l 100 -p 20 -C -M 0.9) were used to detecte the long terminal repeated sequences (LTRs). LTR retriever v2.9.045 with default parameters was used to integrate the prediction results of those two software. To obtain the repetitive sequence in the T. carassii genome by homology-based approach, RepeatMasker v4.1.146 with default parameters was used to search the genome sequence for the sequence similar to the known repetitive sequence in the repetitive sequence database RepBase v20170127 (http://www.girinst.org/repbase). We employed RepeatModeler 2.0.2a47 (-engine ncbi -pa 60 -database) for de novo identification of the remaining repetitive elements in the masked genome. Repeat annotation revealed 25.87 Mb of repetitive sequences, comprising 53.29% of the genome. These repetitive elements primarily consist of interspersed repeats, dominated by two major classes of transposable elements: retrotransposons (23.45%) and DNA transposons (0.15%), alongside a substantial proportion of unclassified repetitive sequences (29.69%) (Table 8).Table 8Statistics of repetitive sequences.ClassSurperFamilyClassificationNumberLength (bp)Percentage in the genomeClassI(Retrotransposon)LTRLTR15851,668,5843.44%LTR/Caulimovirus1210,3410.02%LTR/Copia2937,0480.08%LTR/Gypsy5124,5420.05%LTR/Pao6816,5700.03%LTR/Viper2,9633,822,5087.87%Non-LTRLINE5922,7750.05%LINE/CRE6529,5660.06%LINE/I4,2375,755,08011.85%ClassII(DNA_Transposon)DNA-TEDNA1573,8820.15%OthersUnknownUnknown14,12114,414,30529.69%Total23,20525,875,20153.29% Table 8Statistics of repetitive sequences.ClassSurperFamilyClassificationNumberLength (bp)Percentage in the genomeClassI(Retrotransposon)LTRLTR15851,668,5843.44%LTR/Caulimovirus1210,3410.02%LTR/Copia2937,0480.08%LTR/Gypsy5124,5420.05%LTR/Pao6816,5700.03%LTR/Viper2,9633,822,5087.87%Non-LTRLINE5922,7750.05%LINE/CRE6529,5660.06%LINE/I4,2375,755,08011.85%ClassII(DNA_Transposon)DNA-TEDNA1573,8820.15%OthersUnknownUnknown14,12114,414,30529.69%Total23,20525,875,20153.29% Statistics of repetitive sequences. Gene structure prediction refers to the identification of gene structures such as open reading frames (ORFs), introns and exon regions in genomic sequences. AUGUSTUS v3.5.048 (--species--alternativesfrom-evidence--alternatives-from-sampling--allow_hinted_splicesites = atac), SNAP v6.049 with default parameters, GlimmerHMM v3.0.450 with default parameters, and GeneMark-ET51 with default parameters were used to predict gene structure in the repeat-masked genome. GeMoMa v1.952 with default parameters was used to perform homology prediction and then obtain exon and intron boundary information based on the comparison between the transcript and the genome. EvidenceModler53 was used to integrate the above prediction results. The UTR and other variable cut annotation was predicted by PASA v2.5.254 (-c alignAssembly.config -C -R --ALIGNERS blat, gmap --CPU 32 --stringent_alignment_overlap 30.0). A total of 11,584 genes were predicted, with an average gene length of 1,766 bp. Each gene contains only one transcript isoform, and the vast majority of genes are intronless (only one intron was detected), with an average of one exon per transcript. The corresponding total CDS length is 19,033,979 bp, with an average CDS length of 1,643 bp (Table 2). This gene structural feature indicates a high gene density and a compact genome organization with scarce introns, consistent with the genomic characteristics of Trypanosome. Non-coding RNAs (ncRNAs) refer to RNAs that are not translated into proteins, such as rRNA, tRNA, snRNA, and miRNA. Although these RNAs do not encode proteins, they all play crucial biological roles. The tRNA genes were predicted using tRNAscan-SE v2.0.1155 with default parameters, whereas other ncRNAs were annotated by performing BLAST searches against the database of Rfam v14.3. There are 100 rRNAs, 122 tRNAs, 2 snRNAs, and 322 snoRNAs in the T. carassii genome. After obtaining the genes through gene structure prediction, it is necessary to compare the predicted genes with the functional database to obtain functional information such as protein motifs, domains, functions, and the biological pathways in which they are located. All protein-coding genes were aligned to three integrated protein sequence databases: NR v202108, SwissProt v1.7.1, eggNOG v5. Protein domains were annotated by InterPro v 5.59–91.0. Gene Ontology (GO) terms for each gene were obtained from the corresponding eggNOG-mapper v2.1.12 (-d euk --evalue 1e-5 -itype proteins --override) annotation entry. The pathways in which the genes might be involved were assigned by BLAST56 against the KEGG databases. The protein-coding gene functional annotation results were merged from the above methods. Of the 11,584 protein-coding genes, a total of 11,046 (95.36%) were functionally annotated. Among the individual databases, NR yielded the highest number of annotated genes, with 11,033 (95.24%), followed by eggNOG, which successfully annotated 7,423 genes (64.08%). SwissProt and KEGG annotated 4,852 (41.89%) and 3,642 (31.44%) genes, respectively. GO annotated relatively fewer genes, with 785 genes accounting for 6.78% (Fig. 2C and Table 2). EggNOG functional classification revealed that the predominant biological roles of T. carassii genes are involved in carbohydrate transport and metabolism, signal transduction mechanisms, posttranslational modification, protein turnover, chaperones, translation, ribosomal structure and biogenesis, intracellular trafficking, secretion, and vesicular transport, and chromatin structure and dynamics (Fig. 2D). Similarly, GO enrichment analysis indicated that the major functional categories include metabolic process, biological regulation, response to stimulus, catabolic process, catalytic activity, hydrolase activity, peptidase activity, binding, pyrophosphatase activity, transferase activity, GTPase activity (Fig. 2E). Whole-genome duplication analysis Synonymous substitution rates (Ks) distribution peak finding is a common method to identify whole-genome duplications (WGD) signal. We used blastp to compare the longest protein sequences of genes in the genome, then used DAGchainer57 to identify homologous regions. Finally, we employed the existing script synonym_calc.py3 (https://github.com/tanghaibao/bio-pipeline/blob/main/synonymous_calculation/synonymous_calc.py) to calculate the Ks values for each gene pair within the homologous regions and plotted a density distribution graph. In our study, we used changes in the Ks among paralogous genes (within-genome comparisons) to assess gene duplication and loss over the course of genome evolution. A sudden increase in Ks at a certain period (appearing as a peak on the plot) indicates a large-scale gene duplication event, which likely corresponds to a WGD. Conversely, a smooth decline in Ks suggests gene loss. Notably, the peak (Ks≈0.0176) indicated a recent burst of gene duplication in the genome of T. carassii, which likely corresponds to a WGD (Fig. 2F). The recent WGD event revealed by Ks analysis (Ks≈0.0176) provides a compelling mechanistic explanation for the high gene duplication rate of 16.9% observed in the BUSCO assessment (Table 7). These duplicated genes are likely the retained “fossil” copies resulting from the WGD. During genome survey, the observed heterozygosity of 1.17% (Fig. 1A) may partially reflect the sequence divergence between the two subgenomes that has not yet fully converged following the WGD event. Alternatively, it could be because the WGD event and the origin of this heterozygosity are distinct—meaning that the 1.17% heterozygosity primarily represents polymorphism in the context of population genetics, with the WGD event being superimposed on this background. Data Records The datasets generated in this study are deposited in public repositories to ensure accessibility and long-term preservation. All raw sequencing data generated in this study are available under NCBI BioProject PRJNA134760658, comprising PacBio HiFi reads (SRR3589670059) for primary genome assembly, Hi-C data (SRR3589669860) for chromatin scaffolding, WGS data (SRR3589669961), and RNA-seq data (SRR3589669762) for genome survey. The final chromosome-scale genome assembly is available under GenBank accession GCA_056247815.163. Comprehensive annotation files including structural annotations in GFF3 format and genomic sequences in the fasta format are provided via the Figshare repository (https://doi.org/10.6084/m9.figshare.30103480)64. Technical Validation Genome assembly accuracy was assessed by aligning HiFi reads to the assembled genome using Minimap software36, achieving a 99.77% mapping rate and 100.00% genome coverage, confirming high assembly continuity. Genome completeness was assessed using BUSCO41 with the euglenozoa_odb10 dataset, achieving 100.00% complete genes (83.1% single-copy, 16.9% duplicated), indicating high completeness in the trypanosome genomes that have been sequenced. These results collectively demonstrate the genome’s high continuity, accuracy and completeness, establishing its status as a reference-level genome.