Near complete telomere-to-telomere level genome assembly of blackchin tilapia (Sarotherodon melanotheron)

Abstract

The blackchin tilapia (Sarotherodon melanotheron) is a euryhaline teleost, renowned for its exceptional osmoregulatory ability to thrive across a wide salinity range. However, studies on the genome of Sarotherodon melanotheron are limited, impeding the investigative work on genetic selection at the molecular level. With the sequencing technologies, we present a near complete telomere-to-telomere level genome assembly for Sarotherodon melanotheron, which spans 996.95 Mb with a scaffold N50 of 42.04 Mb and is anchored to 22 chromosomes. This genome assembly was generated by integrating multiple sequencing technologies, including short reads, PacBio HiFi long reads, ONT ultra-long reads, and Hi-C data. The genome assembly demonstrated a high degree of completeness, with 98.36% complete BUSCOs were successfully identified. A total of 391.75 Mb repeating elements and 29,328 protein-coding genes were detected through genome assembly. Functional annotation was successfully completed for 92.10% of the 29,328 protein-coding genes, with a BUSCO score of 99.22%. This genome assembly of Sarotherodon melanotheron not only provides critical insights into its biological traits and genetic diversity but also paves the way for future genetic improvement.

Background & Summary

The blackchin tilapia (Sarotherodon melanotheron), a euryhaline fish native to West Africa, exhibits remarkable osmoregulatory abilities, enabling it to thrive in salinities up to 3-4 times that of seawater¹. In its natural range, it inhabits freshwater, coastal marine, and estuarine environments from Angola to southern Mauritania², where it plays a significant role in both capture fisheries and small-scale aquaculture, contributing to the livelihoods of local communities in West Africa. Due to its strong salinity adaptability, Sarotherodon melanotheron has been introduced to China as a promising genetic resource³ to help overcome the challenge of limited freshwater resources and promote sustainable aquaculture development⁴. Consequently, efforts are now needed to enhance the quality of available germplasm and advance selective breeding methods to support the development of fast-growing strains of Sarotherodon melanotheron, which would benefit the brackish-water aquaculture.

A wealth of research efforts have been dedicated to different field of Sarotherodon melanotheron, including life history^5,6, reproductive biology⁷, genetic diversity^8,9, ecology¹⁰ and osmoregulatory mechanisms^2,11,12. However, owing to the lack of a high-quality reference genome, genetic and functional genomic studies on Sarotherodon melanotheron, has long been stagnant, forcing current research to rely on genomic data from closely related species, such as the Nile tilapia (Oreochromis niloticus). Thus, a high-quality genome assembly for Sarotherodon melanotheron is urgently needed.

In this research, a telomere-to-telomere (T2T) reference genome of the Sarotherodon melanotheron was assembled by integration of Illumina short-sequence reads, PacBio (Pacific Biosciences) HiFi long reads, Hi-C (High-through chromosome conformation capture) sequencing data and ONT (Oxford Nanopore Technologies) ultra-long reads. The assembled genome spans a total length of 996.95 Mb, with a scaffold N50 of 42.04 Mb. The Sarotherodon melanotheron genome contains 29,328 protein-coding genes, 92.10% of which have been functionally annotated. This high-quality genomic dataset is expected to support on species evolution and function research, and molecular breeding of Sarotherodon melanotheron, thereby enhance its economic and ecological values.

Methods

Sample collection

A one-year-old male Sarotherodon melanotheron individual (Fig. 1A) was collected from the Freshwater Fisheries Research Center (FFRC), the Chinese Academy of Fishery Sciences (Wuxi, Jiangsu). Fresh muscle tissues were sampled for whole genome sequencing. A total of seven distinct tissue types, including muscle, brain, gill, liver, spleen, kidney, skin, were meticulously dissected and procured for transcriptome sequencing. Specifically, the fish was euthanized with 100 mg/L MS-222. Target tissues were then rapidly dissected using sterile scissors and forceps, immediately snap-frozen in liquid nitrogen, and stored at −80 °C before usage. This study was approved by the Bioethical Committee of FFRC, Chinese Academy of Fishery Sciences (Approval No. 2013863BCE) and all methods were performed in accordance with the relevant guidelines and regulations of the Declaration of Helsinki. The study was carried out in compliance with the ARRIVE guidelines (https://arriveguidelines.org).

Fig. 1.

The genome survey, Hi-C interaction heatmap, and the Contig distribution map of Sarotherodon melanotheron. (A) A photo of the sequenced fish. (B) A GenomeScope k-mer plot. (C) Hi-C interaction heat map for genome assembly of Sarotherodon melanotheron. (D) Distribution maps of contig on chromosomes with annotated telomere locations of Sarotherodon melanotheron. Contig number: Contig count contained within a chromosome.

Genomic DNA extraction and genome sequencing

High-molecular-weight genomic DNA was extracted using the cetyltrimethylammonium bromide (CTAB) method¹³. Quality assessment of the extracted DNA was evaluated with a NanoDrop 2000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA) and its integrity was further verified by 0.8% agarose gel-electrophoresis. High-quality genomic DNA were submitted to Pisino Biotechnology Co., Ltd. (Nanjing, China) for construction and sequencing.

For short-read sequencing, the library was prepared by TruSeq DNA PCR-free prep kit, followed by sequencing on Illumina NovaSeq^tm X plus platform (Illumina, MA, USA). After removing reads containing adapter, low-quality and ploy-N, a total of 113.70 Gb of clean data were generated. In parallel, the PacBio HiFi sequencing library was built using the PacBio Template Prep Kit 1.0 and sequenced using a PacBio Revio sequencing platform¹⁴. Circular Consensus Sequencing (CCS), implemented in SMRT Link v11.0) software¹⁵, was employed to produce high-accuracy consensus reads, which resulted in the acquisition of ~29.41 Gb of HiFi data.

Ultra-long Oxford Nanopore technologies (ONT) sequencing library was constructed by using the Oxford Nanopore SQK-ULK001 kit, followed by sequencing performed on a PromethION platform. Subsequent processing with NECAT v200221 with default parameters yielded 70.12 Gb of ultra-long reads¹⁶.

Prepared in accordance with Hi- C library protocol¹⁷, the Hi- C library was subjected to sequencing on an Illumina NovaSeq^tm X plus platform (Illumina, MA, USA). In total, approximately 112.68 Gb of Hi-C data were generated.

RNA extraction and full-length transcriptome sequencing

Total RNA extraction was performed via a combined protocol integrating TRIzol reagent with an RNAprep Pure Tissue Kit (TIANGEN Biotech). The SMARTer PCR cDNA Synthesis Kit (Clontech) was used to convert RNA into cDNA through reverse transcription. Subsequently, the full-length RNA-seq library was constructed using the SMRTbell^TM Template Prep Kit 1.0 (Clontech), followed by sequencing on the PacBio Revio sequencing platform with the Sequel^TM Sequencing Kit 2.0. A cumulative dataset of 3.59 Gb of full-length RNA-seq data were generated, providing transcriptomic resources for genome annotation.

Genome assembly and telomere identification

The k-mer analysis was utilized to survey the genome features of Sarotherodon melanotheron with the Illumina short reads using Jellyfish (v2.3.0)¹⁸, followed by analysis of genome survey with GenomeScope (v2.0)¹⁹. Accordingly, the 19-mer analysis was conducted to estimate genomic characteristics of Sarotherodon melanotheron, yielding a predicted genome size of ~1.00 Gb along with a heterozygosity rate of 0.48% (Fig. 1B).

Hifiasm (v0.19.5)²⁰ was employed as the primary assembly tool for PacBio HiFi reads, and the entire process was executed under the software’s default parameter regime. The preliminary assembly was then processed with purge-dups (1.2.5)²¹ to eliminate redundant sequences, resulting in a primary assembly of ~1.19 Gb with a contig N50 of 2.19 Mb.

The HiC-Pro (v3.1.0)²² pipeline was employed to detect valid contact paired reads. The valid reads were processed through the 3D-DNA (parameters: ‘-r 0’)²³ pipeline to anchor the contigs onto chromosomes. To address the existing sequence gaps, ultra-long ONT reads were integrated to generate gapless scaffolds via TGS-GapCloser v1.1.1²⁴ software. To improve the accuracy, genome polishing was performed via Nextpolish2 (0.2.2)²⁵. Consequently, we generated the final genome assembly of Sarotherodon melanotheron, which spanned a total size of 996.95 Mb and is anchored into 22 chromosomes (Fig. 1C,D and Table 1). Its scaffold N50 reaches 42.04 Mb. By performing a genome-wide search with tidk (0.2.0) software²⁶, We additionally detected the distribution and frequency of characteristic sequences (AAACCCT/AGGGTTT) in the telomere region. Ultimately, we identified that the chromosomes of Sarotherodon melanotheron possessed 22 centromeres and 28 telomeres (Table 2).

Table 1.

Statistics of sequencing data for Sarotherodon melanotheron genome assembly and annotation.

Sequencing Library Category	Raw Sequencing data (Gb)	Filtered High-Quality Data (Gb)	Reads Mean Length (bp)	Reads N50 (bp)	Coverage
Short reads Seq	117.23	113.70	150	/	117.59×
PacBio HiFi	29.41	/	17,006	17,489	29.50×
Hi-C	112.68	111.21	150	/	113.02×
ONT	70.12	/	96,738	100,000	70.33×
RNA-seq	3.59	/	2391	2587	/

Table 2.

Positions of telomere and centromere in the genome assembly of Sarotherodon melanotheron.

ID	Sequence Length (bp)	Gaps	Centromere		Telomere
			Start	End	Upstream_start	Upstream_end	Downstream start	Downstream end
chr1	43,199,739	0	7,144,231	7,927,647	1	6,420	43,194,000	43,199,739
chr2	38,260,850	0	4,899,747	5,255,485			38,257,740	38,260,850
chr3	40,418,258	0	4,766,282	5,721,992			40,412,220	40,418,258
chr4	42,462,704	0	35,562,982	35,934,491	1	5,280
chr5	46,711,762	0	5,110,814	5,395,957	1	5,580	46,708,260	46,711,762
chr6	66,404,432	0	11,588,952	12,341,885			66,398,760	66,404,432
chr7	33,122,763	0	29,151,010	29,559,986	1	5,640
chr8	40,879,768	1	32,776,741	33,810,785	1	7,620
chr9	36,122,824	0	7,363,347	8,498,578	1	6,420	36,117,840	36,122,824
chr10	40,620,852	0	5,540,840	5,867,869			40,614,180	40,620,852
chr11	42,862,210	0	3,827,087	6,959,636	1	5,880	42,858,660	42,862,210
chr12	35,091,674	0	28,864,758	29,102,502	1	5,040
chr13	42,909,168	0	8,759,869	9,180,329			42,903,480	42,909,168
chr14	42,042,471	0	28,969,239	29,558,050	1	6,300	42,036,240	42,042,471
chr15	39,569,484	0	36,909,899	39,569,484	1	6,180
chr16	40,803,323	0	40,094,076	40,803,323	1	6,540
chr17	41,030,710	0	4,341,130	5,054,290	1	5,580	41,024,100	41,030,710
chr18	35,514,925	0	6,061,678	6,813,910	1	4,800	35,509,740	35,514,925
chr19	41,544,819	1	16,161,476	16,706,447	1	6,060
chr20	42,410,651	0	4,310,990	4,525,240	1	6,480
chr21	46,688,682	0	4,083,033	4,164,081	1	4,980
chr22	118,277,041	0	13,322,249	13,892,378

Genome annotation

A comprehensive annotation pipeline for repetitive sequences was established, combining de novo prediction and homology-based strategies to detect repetitive sequences in the Sarotherodon melanotheron genome. For the de novo prediction, a custom repeat library was constructed via RepeatModeler v2.0.4²⁷, a computational tool which can initiate two de novo repeat finding programs of RECON v1.0.8²⁸ and RepeatScout v1.0.6²⁹. Afterwards, the homology-based alignment was performed utilizing the RepBase database (v20150807) by RepeatMasker v4.1.4³⁰. A total of 1,501,443 repetitive elements were identified, constituting 39.3% of the entire assembled genome (Table 3). Among these repetitive elements, DNA transposons were the most prevalent at a proportion of 17.59%, followed by long interspersed elements (LINEs, 9.69%), long terminal repeats (LTRs, 4.21%), and Rolling-circles (1.33%) (Fig. 2, Table 3).

Table 3.

Statistics of repetitive sequences in Sarotherodon melanotheron.

Elements	Elements Count	Length (bp)	Genome Percentage (%)
Retroelements/SINEs	41,422	5,659,241	0.57
Retroelements/LINEs	279,225	96,623,271	9.69
Retroelements/LTRs	124,949	41,963,614	4.21
BEL/Pao	7,148	3,320,577	0.33
Ty1/Copia	5,739	1,885,174	0.19
Gypsy/DIRS1	36,725	14,141,033	1.42
others	75,337	22,616,830	2.27
DNA transposons	718,204	175,328,581	17.59
hobo-Activator	202,843	57,798,324	5.80
Tc1-IS630-Pogo	191,325	57,358,889	5.75
MULE-MuDR	4,022	327,517	0.03
PiggyBac	17,476	3,381,431	0.34
Tourist/Harbinger	24,762	6,179,445	0.62
others	277,776	50,282,975	5.05
Rolling-circles	29,605	13,286,822	1.33
Unclassified	300,459	56,436,253	5.66
Satellites	5,803	2,413,521	0.24
Simple repeats	1,776	73,855	0.01
Total repeats	1,501,443	391,785,158	39.3

Fig. 2.

Features of Sarotherodon melanotheron genome assembly. Tracks from outer to inner represent (a) Length of chromosome; (b) GC content distribution; (c) Density of genomic code gene; (d) Density of repetitive sequence; (e) Density of long terminal repeat (LTR) element; (f) Density of Gypsy elements; (g) Density of Copia elements; (h) Density of DNA transposon; (i) The interconnections within the circle represent syntenic gene pairs.

Gene prediction and annotation. To systematically annotate protein-coding genes, a combination of three approaches including de novo gene prediction, homology-based annotation, and transcriptome-based annotation was applied in this research. The de novo predictions of protein-coding genes were performed via four software tools, Augustus(v3.3.2)³¹, SNAP(v2017-03-01)³², GlimmerHMM (v3.0.4)³³ and GeneMark (v4.71)³⁴. And for homology-based prediction, we downloaded the protein sequences of 5 closely related representative fish species from NCBI, including Oreochromis niloticus, Simochromis diagramma, Oreochromis aureus, Astatotilapia calliptera and Haplochromis burtoni. After acquiring genomic data, sequence alignment was performed against the Sarotherodon melanotheron T2T genome using GeMoMa v2.3³⁵ and Exonerate (v2.2.0)³⁶. For transcriptome-based prediction, the de novo assembly of full-length RNA-Seq datasets was performed using the software Trinity (r20140717)³⁷. Subsequently, the gene structure of the assembled transcripts were annotated with the software PASA (2.5.2)³⁸. Ultimately, a comprehensive gene annotation was implemented in EvidenceModeler (r2012-06-25), which integrated the results of de novo gene predictions, homology-based predictions, and transcriptome-based predictions³⁹. Consequently, 29,328 protein-coding genes were successfully identified, with the average length of individual gene reaching 18,699.6 bp and the average length of transcription sequence measuring 1,647.4 bp. According to the statistical analysis, the average number of exons per gene was 9.58, with exons and introns having average length of 171.9 bp and 1,987 bp respectively (Table 4). Comparative analyses of the statistical characteristics of gene set elements between Sarotherodon melanotheron and other five closely related species revealed similar distribution patterns (Fig. 3).

Table 4.

Statistics of predicted protein-coding genes in Sarotherodon melanotheron.

Statistic	Value
Total length of genes (bp)	548,424,058
Genes proportion of genome (%)	55.01
Total number of genes	29,328
Average length of gene (bp)	18,699.6
Total exon number	281,016
Average exon length (bp)	171.9
Average length of introns	1,987
Total length of CDSs	48,316,575
Proportion of CDSs in the genome (%)	4.85
Average transcription length (bp)	1,647.4

Fig. 3.

Comparisons of the statistical characteristics of the gene set elements between Sarotherodon melanotheron, Oreochromis niloticus, Simochromis diagramma, Oreochromis aureus, Astatotilapia calliptera and Haplochromis burtoni. (A) Distributions of gene length. (B) Distributions of CDS length. (C) Distributions of exon length. (D) Distributions of intron length distributions.

To conduct functional annotation of the predicted protein-coding genes, the constructed gene set was aligned against publicly accessible functional databases. More specifically, sequence similarity search (with an E-value threshold 1e-5) was conducted via DIAMOND blastp (2.0.14.152)⁴⁰ against the NCBI NR (Non-Redundant Protein Sequence) and SWISS-Prot databases. GO (Gene Ontology) and Pfam (Protein Family Database) annotations were generated via InterProScan (5.61–93.0)⁴¹, which was applied to perform domain similarity searches against respective functional database. In addition, the KEGG Automatic Annotation Server (KAAS)⁴² was used for KEGG (Kyoto Encyclopedia of Genes and Genomes annotation, with the aim of mapping these predicted protein-coding genes to their relevant KEGG pathways. As a consequence, 27,010 genes (92.10% of the predicted genes in Sarotherodon melanotheron were annotated in at least a single functional database (Table 5).

Table 5.

Functional annotation statistics of identified protein-coding genes in Sarotherodon melanotheron.

	Database	Number	Percent (%)
Total	/	29328	100
Annotated	NR	26998	92.06
	Swiss-Prot	22910	78.12
	Pfam	23096	78.75
	KEGG	17729	60.45
	GO	18434	62.85
Unannotated	/	2318	7.90

Non-coding RNA prediction and annotation. tRNA genes annotation across the whole gene was accomplished with, TRNAscan-SE (1.3.1)⁴³, whereas rRNA gene prediction was performed with RNAmmer (1.2)⁴⁴.For other non-coding RNAs prediction, prediction analysis mainly relied on sequence aliganment against the Rfam database (1.0)⁴⁵. A total of 5,127 non-coding RNAs were identified with average length of 111.80 bp and cumulative length of 573,201 bp, accounting for approximately 0.0575% of the entire genome (Table 6).

Table 6.

Annotation statistics of identified non-coding RNA in Sarotherodon melanotheron.

Type	Copy number	Average length (bp)	Total length (bp)	Percentage of genome (%)
18 s rRNA	17	1,861.12	31,639	0.0032
28 s rRNA	20	4,782.70	95,654	0.0096
tRNA	3,298	75.32	248,404	0.0249
others	1,792	110.21	197,504	0.0198

Data Records

The Illumina short-read sequencing data were deposited in the Genome Sequence Archive (GSA) at the National Genomics Data Centre (NGDC), part of the China National Centre for Bioinformation (CNCB)/Beijing Genomics Institute, Chinese Academy of Sciences, under the accession number CRR2424486⁴⁶. The PacBio HiFi sequencing data were deposited in GSA at NGDC under the accession number CRR2424485⁴⁷. The Hi-C sequencing data were deposited in GSA at NGDC under the accession number CRR2424487⁴⁸. The ONT ultra-long sequencing data were deposited in GSA at NGDC under the accession number CRR2424484⁴⁹. The full-length transcriptome sequencing data were deposited in GSA at NGDC under the accession number CRR2424488⁵⁰. The assembled genome sequences have been deposited in the European Nucleotide Archive (ENA) under the accession number GCA_982250375.1⁵¹. The genome annotation files have been deposited at Figshare⁵².

Technical Validation

BUSCO, v5.4.5⁵³, coupled with arthropoda_odb10 database were employed to evaluate the completeness and accuracy of Sarotherodon melanotheron genome assembly. The BUSCO evaluation results demonstrated that 3299 (98.36%) BUSCOs were identified as complete, among these, 3269 (97.47%) were classified as single-copy and 30 (0.89%) as duplicated ones, respectively (Table 7). These findings confirm a good quality of the genome. Chromosome numbers of the Sarotherodon melanotheron genome were confirmed by the Hi-C (Fig. 1C).

Table 7.

BUSCO assessment results of the Sarotherodon melanotheron genome assembly and annotation.

Item	Assembly		Annotation
	Number	Percent (%)	Number	Percent (%)
Count of Complete BUSCOs (C)	3,299	98.36	3,328	99.22
Fully matched single-copy BUSCOs (S)	3,269	97.47	3,268	97.44
Fully matched duplicated BUSCOs (D)	30	0.89	60	1.79
Fragmented BUSCOs (F)	17	0.51	2	0.06
Unidentified BUSCOs (M)	38	1.13	24	0.72
Total BUSCO groups searched	3354	100	3354	100

An addition BUSCO assessment was performed to evaluate the annotation quality of protein-coding genes and revealed 99.22% completeness, with 97.44% single-copy, 1.79% duplicated, and 0.06% fragmented orthologs, confirming high-quality gene annotation.

Author contributions

J.Q. and Y.T. conceived and designed the project. S.L. and Y.D. collected the experimental samples for this study. W.W. and J.H., performed the genome assembly and conducted bioinformatics analysis. C.Z. and Y.T. drafted the manuscript. J.Q., P.X. and W.T., provided detailed guidelines for manuscript refinement and made revisions to the manuscript. All authors have read the final manuscript carefully and endorsed the final draft of the manuscript.

Data availability

The genome sequencing data, including Illumina short-read, PacBio HiFi, Hi-C, and ONT ultra-long, can be publicly available from the NGDC GSA database under the accession number CRR2424486, CRR2424485, CRR2424487, and CRR2424484. The full-length transcriptome sequencing data used for genome annotation also can be publicly obtained from the NGDC GSA database under the accession number CRR2424488. The genome assembly can be publicly available from the ENA under accession number GCA_982250375.1. The genome annotation files can be publicly downloaded at Figshare (10.6084/m9.figshare.31282306.v2).

Code availability

All scripts and pipelines applied to genome assembly and gene annotation strictly followed the standard manuals and protocols of the utilized bioinformatics software. Notably, no specific code was developed for this study.

Competing interests

The authors declare no competing interests.