Background & Summary

Alectoris Chukar (chukar) commonly known as “chukar”, is a member of the Galliformes order and the Phasianidae family, hailing from the stony semi-desert regions of Asia, Western Europe, and the Middle East. This species has been introduced to numerous other countries such as the United States, Canada, England, and New Zealand, primarily for stocking on game farms or releasing for hunting purposes1,2. In recent years, there has been an uptick in the use of chukars for meat production under controlled husbandry conditions. Owing to their rapid growth, high productivity, and superior meat quality, chukars are ideally suited for commercial production3,4,5,6. The domestication of these partridges coupled with selection for growth traits enhances their potential as a prime source of high-quality protein for human consumption.

The wild chukar, a polytypic species with 22 subspecies scattered globally, exhibits a broad spectrum of environmental adaptations. Among these are the six subspecies of pubescens, potanini, pallida, falki, dzungarica, and pallescens, with pubescens and pallida exclusively found in China. These subspecies have evolved to survive in their specific habitats, which span a wide range of temperatures and altitudes. For instance, while falki is adapted to a drier environment, pallescens thrives in the highlands of Tibet (altitude >4000 m).

This environmental adaptation capacity and genetic diversity among the chukar subspecies underline the significance of comprehensive genomic research in this species. Whole-genome sequence assembly has proven to be a fundamental tool for extensive genomics initiatives, including evolutionary studies and efficient breeding strategies. Over the years, critical poultry species like chickens, turkeys, and ducks have substantially reaped the benefits of these genomic resources7,8,9,10. Notably, the chicken reference genome has undergone several refinements, making it one of the superior vertebrate genomes available and establishing it as a model for avian research9. However, the chukar partridge’s genomic advancement is currently stymied due to the absence of a reference genome. The introduction of phased-genome assembly, renowned for its precision in resolving complex genomic variations11, could be instrumental in breaking this impasse, thus transforming the genetic improvement of chukar breeding, spurring evolutionary studies, and securing genetic resource conservation.

In this study, we employ a de novo assembly strategy to present the first continuous, accurate, phased-resolved genome for the chukar. Utilizing this genome, we resequenced and analyzed five wild subspecies and one domestic population of chukar. Our research provides valuable resources for investigating adaptive evolution, breeding, and conservation genomics of crucial ecological species.

Method

Ethics statement

The collection and handling of the samples in this study were carried out in accordance with approved guidelines and regulations from both Lanzhou University and Shanghai Jiao Tong University.

Sample collections

We sequenced the genome of a female domesticated chukar collected from Tianming Alectoris chukar Farm (Guangzhou, Guangdong, China), which was primarily used for genome assembly. The transcriptomes of two adult female domesticated chukars, which were collected from Tianming Alectoris Chukar Farm (including the one used for genome assembly), were sequenced for annotation of coding genes in the genome. In addition, we included the genomes resequenced from a total of 58 chukars for genotype identification. These samples include 14 pubescens, 12 potanini, 9 pallida, 14 falki, 2 pallescens, and 7 domestic chukars (refer to Table 1 for more details). Pubescens, potanini, pallida, and falki muscle samples were obtained from 10 distinct locales including Akesai(AK), Changji(CJ), Dongdashan(DD), Helanshan(HL), Jingtai(JT), Kuerle(KE), Quzi (QZ), Subei (SB), Tongchuan (TC) and Wudu (WD) during 2002–2008, representing the majority of chukar’s geographical area in China and reflecting various geographic, topographic, and climatic conditions (Fig. 1, Table 1). To avoid sampling near relatives, each bird within a location was collected from a different portion of the colony. These 49 samples were donated by Lanzhou University. Three and four blood samples were collected at random from domestic (DOM) chukars in Tianming farm (Jieyang, China) and Qinxiangyuan farm (Jieyang, China), respectively. To avoid the selection of relatives, the pedigrees of these DOM chukar were investigated (Table 1). Two muscle samples of pallescens sampled in Tibet (TB) province were received from the animal branch of the southwest China germplasm bank of wildlife (Yunnan, China) (Table 1).

Table 1 Sample sizes and sampling locations of chukar subspecies.
Full size table
Fig. 1
figure 1

Sampling distribution map of chukar subspecies. Pubescens (orange), potanini (red), pallida (green), falki (blue), pallescens (brown), and domestic (pink) samples were obtained from 12 distinct locales including Akesai(AK), Changji(CJ), Dongdashan(DD), Helanshan(HL), Jingtai(JT), Kuerle(KE), Quzi (QZ), Subei (SB), Tongchuan (TC),Wudu (WD), Tibet (TB), and Jieyang(DOM). The horizontal axis represents longitude, and the vertical axis represents latitude.

Full size image

De novo sequencing and assembly of the chukar phased genome

The DNA samples of the female domesticated chukar blood were extracted and sequenced using PacBio single-molecule real-time (SMRT) sequencing and Illumina paired-end sequencing technology. We carried out SMRT DNA sequencing of ∼20 kb inserts using the PacBio Sequel II platform (Personal Biotechnology Co., Ltd. Shanghai, China). Next, 400 bp paired-end libraries (refer to Illumina TruSeq DNA Sample Preparation Guide) constructed from the same genomic DNA were sequenced on the Illumina HiSeq platform (Personal Biotechnology Co., Ltd. Shanghai, China). The DNA sample of the chukar muscle was extracted and sequenced using Hi-C technology in Beijing Nuohezhiyuan Technology Service Co, Ltd. We filtered and trimmed the Illumina and Hi-C reads to remove adapters and low-quality bases using the standard settings in SOAPnuke v1.5.012.

After low-quality and adaptor reads were filtered, we obtained ~136.59 Gb long sequencing data that were used to assemble the genome (Table S1a). De novo assembly followed the PacBio string graph assembler process, using FALCON v2.1.4, FALCON-Unzip v0.4.0, and FALCON-phase v0.2.013 to generate long-range phased haplotypes. FALCON computes an initial assembly by correcting errors in raw reads and subsequently assembling them using a string graph formulated from read overlaps. Following this, FALCON-Unzip identified read haplotypes based on the phasing information derived from detected heterozygous positions. These phased reads are then deployed to assemble both haplotigs and primary contigs. As a result of this assembly process, there were 363 and 1,711 contigs for the primary contigs and haplotigs respectively (Table 2). The cumulative lengths of these two haploid assemblies were 1.03 Gb and 0.85 Gb, with their Contig N50 lengths reaching approximately 29.3 Mb and 1.1 Mb each. Subsequently, FALCON-Phase inputs the partially phased long-read assembly from FALCON-Unzip, and extends the phasing on the contigs using ~182 Gb filtered Hi-C data from the same sample. This achieves a phased contig-level genome assembly of chukar, resulting in the production of both Hap1 Contigs and Hap2 Contigs with a total length of 1.03 Gb and contig N50 length of 1.06 Mb (Table 2, Table S1a). Furthermore, Jellyfish v2.1.4 (https://github.com/gmarcais/Jellyfish) was used in conjunction with GenomeScope v1.0.014 to calculate genome size and heterozygosity in the chukar genome a k-mer frequency of 18. The genome size estimated (~1.0 Gb) was consistent with the genome assembled (Fig. 4a).

Table 2 Summary of genome assembly of Alectoris Chukar.
Full size table

After the contig-level assemblies were generated, they were respectively polished using Pilon v1.2215, with the help of ~124.12 Gb high-quality Illumina data. Scaffolding was performed on these polished sequences using ALLHiC v0.9.816, with the support of ~182 Gb Hi-C sequencing data. In our next steps, Ragag Scaffold v1.1.017 was employed for a reference genome-assisted methodology to construct more comprehensive chukar haplotype genomes and ascertain chromosome IDs of scaffolds. The chicken chromosome-assembled genome (GRCg6a)18 served as the reference due to its close evolutionary relationship with chukar and shared chromosome number (n = 78)19. This stage merely oriented and ordered draft assembly sequences into longer sequences without modifying the input query sequence. Finally, we generated two phased pseudo-haplotype genomes of chukar (Hap1 Scaffolds and Hap2 Scaffolds) with 31 pairs of chromosomes (Fig. 2b and Table S2). The N50 length for both Hap1 Scaffolds and Hap2 Scaffolds was significantly enhanced, reaching 93.6 Mb and 93.5 Mb respectively.

Fig. 2
figure 2

Diagram of genome assembly and phylogenetic relationship of Alectoris chukar (chukar). (a) Photograph of the chukar. (b) Synteny map of primary and associated assembly.

Full size image

We evaluated the completeness of two haplotype genomes via BUSCO v5.1.220 benchmarking using the aves_odb10 dataset. BUSCO reported 96.8% and 96.5% complete universal single-copy orthologs and a low rate of duplication (0.3% and 0.5%) for the Hap1 scaffolds and Hap2 scaffolds, respectively (Fig. 3). This indicated that the genome quality of Hap1 scaffolds and Hap2 scaffolds was better than those of other recently published bird genome assembly21,22,23,24. Because of the higher completeness and quality of the scaffolded Hap1 assembly, it was selected as the chukar reference genome for downstream analysis. The pairwise genome alignments of Hap1 and Hap2 were performed using MUMer v4.0.025. The collinearity analysis of the Hap1 scaffolds and Hap2 scaffolds is shown in Fig. 2b.

Fig. 3
figure 3

Histogram of BUSCO assessment of the chukar haplotype genome.

Full size image
Fig. 4
figure 4

(a) The distribution of the k-mer frequency of the chukar genome. (b) Genome-wide Hi-C heatmap of Alectoris chukar, where both the horizontal and vertical axes represent genomic loci along chromosomes. The chromosomes are arranged sequentially from top to bottom (on the vertical axis) and from left to right (on the horizontal axis), displaying chromosomes Chr1 through Chr28, followed by Chr30, Chr31, Chr33, ChrZ, and ChrW. (c)The pairwise genome alignments of the chukar genome and the chicken genome are shown, displaying chromosomes with a length greater than 1 Mb.

Full size image

Repeat sequence and gene annotation

To identify genomic repeats, we used the RepeatMasker v4.1.2 (http://www.repeatmasker.org/) to scan the chukar genome sequence of primary assembly. RMBlast v2.11.0 (https://www.repeatmasker.org/rmblast/) was used as the alignment engine. The Dfam 3.026 and Repbase27 were used and we specified the species library as ‘chicken’ for the chukar. The overall GC content of the chukar genome was estimated to be 41.81%, which is similar to that of the other reference bird species. Interspersed repeats accounted for approximately 9.8% of the whole genome, spanning 101.47 Mb, and consisted of approximately 90.78 Mb retroelements and 10.27 Mb DNA transposons. Approximately 7.19% of the sequences were identified as long interspersed nuclear elements (LINEs), which were thus the largest component, whereas 1.52% of the sequences were identified as long terminal repeats (LTRs). The chicken repeat 1 group was the most abundant, occupying 99.9% of the identified LINEs. The overall level of repetitive content in the chukar (Table 3) was similar to that in the common pheasant21 and chicken18 and greater than those of the turkey7 as well as most sequenced birds28, which may be attributed to the advantage of the long-read sequencing technology.

Table 3 The abundance of repeat elements in chukar genome.
Full size table

Gene annotation was performed using a combined strategy of ab initio predictions, homologue prediction, and transcriptome evidence. RNA from the liver, spleen, muscle, thymus gland, bursa of fabricius, and kidney of these two chukars was isolated, library prepared, and sequenced using Illumina technology in Beijing Nuohezhiyuan Technology Service Co, Ltd. Equal amounts of RNA from each of these six tissues were mixed for single-molecule long-read RNA sequencing (Iso-Seq) to obtain full-length transcriptomic data. Illumina reads were filtered by Trimmomatic v0.3929 and then combined to input to Trinity v2.14.0 (https://github.com/trinityrnaseq/trinityrnaseq) for transcript assembly. Iso-Seq reads were processed using SMRT tools v5.1.0 and ISO-SEQ v3.1 software packages. The pipeline includes five main steps to obtain high-quality sequences: (1) generating circular consensus (CCS) reads, (2) demultiplexing and primer removal and classifying full-length CCS reads, (3) clustering full-length non-chimeric (FLNC) sequences, and, finally, (4) polishing FLNC sequences (Table S1d). We use PASA v2.5.2 (https://github.com/PASApipeline/PASApipeline) to align assembled transcripts obtained from Illumina reads and Iso-Seq reads to the chukar genome sequences and then TransDecoder v5.50 (https://github.com/TransDecoder/TransDecoder) from the Trinity v2.14.0 to identify the likely open reading frame within the transcripts. The ab initio gene prediction was performed using Augustus based on chicken models30. For homology-based annotation, GeMoMa v1.931 was employed using gene annotation information from chicken. Finally, all the results were integrated using the EVidenceModeler pipeline v1.1.132. Alternative splicing analysis of the transcripts uses SUPPA v2.333, which generates seven different alternative splicing types, including skipped exon (SE), alternative 5′/3′ splice sites (A5/A3), mutually exclusive exons (MX), retained intron (RI) and alternative first/last exons (AF/AL). The final gene models comprised 20,082 transcripts, spanning a 302 Mb genomic region. These genes were annotated by using the UniProt/SwissProt protein database and validated 17,997 protein products. In addition, 109,477 splicing events were identified corresponding to all genes. Among seven type splicing events, the largest number of splicing events in chukar was alternative first exons (28,722, 26%), followed by Retained Intron (25865, 24%) and alternative 5′ splice site (17235, 16%) (Table S2b).

Population-based resequencing and variation calling

Genomic resequencing was performed for each individual on MGI-SEQ. 2000 platform. Raw reads were subjected to SOAPnuke v1.5.012 processing to remove sequencing adapters and low-quality reads. Following whole-genome resequencing of domestic chukar and five wild chukar subspecies, a total of 1,071.8 Gb of the clean base was obtained, with an average depth of 18× (Table S1b). High-quality reads were aligned to the chukar genome that we assembled above using the Burrows−Wheeler Aligner v0.5.934. Variants were called using the GATK tool suite v4.2.6.135. Briefly, potential PCR duplicates were marked using MarkDuplicates option. The HaplotypeCaller option was used to construct general variant calling files for all the samples by invoking -ERC:GVCF. All of gVCF files were combined using GenotypeGVCFs option to form a single variant calling file. To obtain high-quality SNPs and Indels, we used the GATK hard filter to filter the merged VCF data with the best practices recommended parameters36. After quality filtering, 2,574,885 high-quality INDELs and 14,988,840 high-quality SNPs were identified. Following the removal of SNPs and INDELs using vcftools v0.1.1637 with the parameters ‘–mac 3–maf 0.05–min-meanDP 5–minQ 30–max-missing 0.95’, a total of 6,991,669 SNPs and 757,682 INDELs from 58 individuals were retained.

Data Records

The chukar genome assembly reported in this paper have been deposited in the Genbank under the project PRJNA780965 with the accession number JAXHPU00000000038. The variation files for this study are located under analysis ERZ22149693 at EVA (European Variation Archive), with the accession number PRJEB713396039. The PacBio sequencing data (SRR27640724) were specifically used to construct the genome assembly. For variant analysis, we utilized whole-genome resequencing data (SRR16961228-SRR16961264). Additionally, transcriptome data from (SRR26796665-SRR26796677 and SRR27640723) informed our genome annotation. These datasets, integral to our project, are available through NCBI BioProject PRJNA78096540.

Technical Validation

Jellyfish was used in conjunction with GenomeScope14 to calculate genome size and heterozygosity in the chukar genome using a k-mer frequency of 18. The genome was consistent with the genome size estimated (1.0 Gb) (Fig. 3a). The Hi-C heatmap revealed a well-organized interaction contact pattern along the diagonals within/around the chromosome (Fig. 3b), which indirectly confirmed the accuracy of the chromosome assembly. The pairwise genome alignments of the chukar genome and the chicken genome using NGenomeSyn41 demonstrate a high level of consistency, which also implies that the genome assembly is accurate and reliable (Fig. 3c). To further assess the quality of assembled chukar genome, we have undertaken an extensive quality assessment by mapping 58 genome resequencing samples to it. The resulting quality metrics, including a coverage of over 99%, a sequencing depth greater than 15X, and a mapping rate of more than 98%, are detailed in Table S3. These metrics not only confirm the technical soundness but also the high quality of the genome assembly. We also evaluated the completeness of two haplotype genomes via BUSCO v5.1.2 benchmarking using the aves_odb10 dataset20. BUSCO reported 96.8% and 96.5% complete universal single-copy orthologs and a low rate of duplication (0.3% and 0.5%) for the Hap1 scaffolds and Hap2 scaffolds, respectively (Fig. 3).