Chromosomal-level genome assembly and single-nucleotide polymorphism sites of black-faced spoonbill Platalea minor

Abstract

Platalea minor, or black-faced spoonbill (Threskiornithidae), is a wading bird confined to coastal areas in East Asia. Due to habitat destruction, it was classified as globally endangered by the International Union for Conservation of Nature. However, the lack of genomic resources for this species hinders the understanding of its biology and diversity, and the development of conservation measures. Here, we report the first chromosomal-level genome assembly of P. minor using a combination of PacBio SMRT and Omni-C scaffolding technologies. The assembled genome (1.24 Gb) contains 95.33% of the sequences anchored to 31 pseudomolecules. The genome assembly has high sequence continuity with scaffold length N50 = 53 Mb. We predicted 18,780 protein-coding genes and measured high BUSCO score completeness (97.3%). Finally, we revealed 6,155,417 bi-allelic single nucleotide polymorphisms, accounting for ∼5% of the genome. This resource offers new opportunities for studying the black-faced spoonbill and developing conservation measures for this species.

Introduction

The black-faced spoonbill Platalea minor (Threskiornithidae) (NCBI:txid259913, Figure 1A) is confined to coastal areas in East Asia, including Hong Kong, Macau, Taiwan, Vietnam, North Korea, South Korea, and Japan. The natural habitats of P. minor have been disturbed by human activities and industrialization, leading to the decline in the bird population over the last century [1, 2]. With an estimation of more than 6,000 individuals worldwide, the International Union for Conservation of Nature (also known as IUCN) has categorised the black-faced spoonbill as a globally endangered species. A quarter of the worldwide population of P. minor can be found in Hong Kong, and it is protected locally under the Wild Animals Protection Ordinance Cap 200. Genetic methods, including studies on genetic diversity and population structure, have been used to help retain this species with high conservation value [3, 4]. Nevertheless, a reference genome of this species was missing.

Figure 1.

(A) Picture of Platalea minor; (B) Statistics of the genome assembly generated in this study; (C) Hi-C contact map of the assembly visualised using Juicebox (v1.11.08); (D) Repetitive elements distribution.

Methods

Sample collection

Tissue samples of 14 P. minor individuals were collected from the north and northwestern parts of the New Territories, Hong Kong, between February 2015 and February 2020, with help from Kadoorie Farm and Botanic Garden. These samples were stored in 95% ethanol. Details of the sample collection are listed in Table 1.

Table 1.

Summary of sequencing data.

Sample	No. of reads	No. of bases	Coverage (×)	Accession number	KFBG reference number	Provenance details	Date specimen acquired at KFBG	Specimen type
Sequencing data for reference genome
PacBio HiFi	2,707,085	25,352,990,722	20	SAMN35152374	K14352	Mai Po	15/02/2020	Tissue in 95% ethanol
Omni-C	518,587,164	77,788,074,600	63	SAMN40731791	B0181-2004	Lok Ma Chau	26/01/2004	Tissue sample to be taken from carcass
Population resequencing data
BFS1	39,968,142	5,995,204,753	4.8	SAMN35319659	K7366	Mai Po fishpond	23/02/2015	Tissue in 95% ethanol
BFS2	41,230,936	6,184,622,630	5.0	SAMN35319660	K7388	Lok Ma Chau fishpond	13/04/2015	Portion of toe in 95% ethanol
BFS3	40,776,316	6,116,429,712	4.9	SAMN35319661	K8627	Ma Cho Lung fishpond	11/02/2016	Tissue in 95% ethanol
BFS4	39,063,262	5,859,473,834	4.7	SAMN35319662	K11131	Mai Po Gei Wai	26/02/2018	Tissue in 95% ethanol
BFS5	44,638,096	6,695,693,640	5.4	SAMN35319663	K11194	Lut Chau, Nam San Wai	10/04/2018	Tissue in 95% ethanol
BFS6	41,890,130	6,283,502,065	5.1	SAMN35319664	K12542	Lok Ma Chau pond	18/12/2018	Blood in 95% ethanol
BFS7	41,319,808	6,197,954,720	5.0	SAMN35319665	K12611	Lok Ma Chau pond	17/01/2019	Tissue in 95% ethanol
BFS8	42,068,480	6,310,254,770	5.1	SAMN35319666	K12706	Lok Ma Chau	03/03/2019	Blood in 95% ethanol
BFS9	42,539,460	6,380,900,668	5.1	SAMN35319667	K14133	Lok Ma Chau	26/11/2019	Tissue in 95% ethanol
BFS10	40,359,264	6,053,871,746	4.9	SAMN35319668	K14208	Lok Ma Chau	25/12/2019	Tissue in 95% ethanol
BFS11	41,496,386	6,224,440,405	5.0	SAMN35319669	K14366	Mai Po	20/02/2020	Portion of feather in 95% ethanol
BFS12	39,361,374	5,904,189,803	4.8	SAMN35319670	K14401	Tin Shui Wai Wetland Park	12/03/2020	Tissue in 95% ethanol
BFS14	41,319,546	6,197,915,065	5.0	SAMN35319671	K14324	Tin Shui Wai Wetland Park	03/02/2020	Portion of feather in 95% ethanol

Isolation of high molecular weight genomic DNA

High molecular weight (HMW) genomic DNA was extracted from a single individual, labelled “BFS13”. The tissue sample was first ground into a powder with liquid nitrogen and then processed with the Qiagen MagAttract HMW kit (Qiagen Cat. No. 67563), following the manufacturer’s protocol. The final DNA sample was eluted with 120 μL of elution buffer (PacBio Ref. No. 101-633-500) and subjected to quality checks using the NanoDrop™ One/OneC Microvolume UV–Vis Spectrophotometer, Qubit^® Fluorometer, and overnight pulse-field gel electrophoresis.

DNA shearing, PacBio library preparation, and sequencing

Approximately 4.4 μg of HMW DNA was processed with DNA shearing through six centrifugation steps in a g-tube (Covaris Part No. 520079) at 2,000 × g for 2 min. The sheared DNA was transferred to a 2 mL DNA LoBind^® Tube (Eppendorf Cat. No. 022431048) and stored at 4 °C. Overnight pulse-field gel electrophoresis was used to assess the fragment size distribution of the sheared DNA. Next, an SMRT bell library was constructed using the SMRTbell^® prep kit 3.0 (PacBio Ref. No. 102-141-700), following the manufacturer’s instructions. Briefly, the sheared DNA was processed with DNA repair, followed by polishing and tailing with A-overhang at both ends of each DNA strand. T-overhang SMRTbell adapters were then ligated to the polished ends to form SMRTbell templates, which were purified with SMRTbell^® cleanup beads (PacBio Ref. No. 102158-300). The quantity and fragment size of the SMRTbell library were inspected with Qubit^® Fluorometer and overnight pulse-field gel electrophoresis, respectively. A nuclease treatment was conducted to remove any non-SMRTbell structures, and a subsequent size-selection step with 35% AMPure PB beads was used to remove short fragments. The final preparation of the library was performed using the Sequel^® II binding kit 3.2 (PacBio Ref. No. 102-194-100). In brief, Sequel II primer 3.2 and Sequel II DNA polymerase 2.2 were added to anneal and bind to the SMRTbell templates, respectively. An internal control provided by the kit was also added. Finally, the library was loaded on the PacBio Sequel IIe System at an on-plate concentration of 90 pM with the diffusion loading mode. The sequencing was run in 30-h movies, with 120 min pre-extension. In total, one SMRT cell was used to output high-fidelity (HiFi) reads, and the sequencing data details are listed in Table 1.

Omni-C library preparation and sequencing

An Omni-C library was constructed using the Dovetail^® Omni-C^® Library Preparation Kit (Dovetail Cat. No. 21005), following the manufacturer’s protocol. A total of 80 mg of tissue was ground into a powder with liquid nitrogen, transferred to 1 mL 1× PBS, and then subjected to crosslinking with formaldehyde and digestion with endonuclease DNase I. An aliquot of 2.5 μL lysate was used for assessing lysate quantification and fragment size distribution using Qubit^® Fluorometer and TapeStation D5000 HS Screen Tape, respectively. Then, end polishing, bridge ligation, and proximity ligation were carried out in the crosslinked DNA fragments. Next, crosslink reversal was performed, followed by DNA purification and size selection with SPRIselect™ Beads (Beckman Coulter Product No. B23317). The library preparation was continued with end repair and adapter ligation using the Dovetail™ Library Module for Illumina (Dovetail Cat. No. 21004), followed by DNA purification with SPRIselect™ Beads. The DNA fragments were then captured with Streptavidin Beads and Universal and Index PCR Primers from the Dovetail™ Primer Set for Illumina (Dovetail Cat. No. 25005) were added to amplify the DNA library. A final size selection was carried out using SPRIselect™ Beads to retain DNA fragments ranging between 350 bp and 1000 bp. The quantity and fragment size distribution of the library were inspected by the Qubit^® Fluorometer and the TapeStation D5000 HS ScreenTape, respectively. The final library was sequenced on an Illumina HiSeq-PE150 platform at Novogene. The details of the sequencing data are listed in Table 1.

Genome assembly and gene model prediction

De novo genome assembly was performed using Hifiasm (RRID:SCR_021069) [5]. Haplotypic duplications were identified and removed using purge_dups (RRID:SCR_021173) based on the depth of HiFi reads [6]. Proximity ligation data from the Omni-C library was used to scaffold genome assembly by YaHS (RRID:SCR_022965) [7]. Transposable elements (TEs) were annotated using the automated Earl Grey TE annotation pipeline (version 1.2) as previously described [8]. Genome annotation was performed using Braker (v3.0.8) (RRID:SCR_018964) [9] with default parameters. Briefly, the genome was soft-masked using redmask (v0.0.2) [10]. A total of 2,468,534 aves reference protein sequences were downloaded from NCBI as protein references. A blood RNA-Seq dataset (SRR6650848) [11] was also downloaded from NCBI and aligned to the soft-masked genome using hisat2 (RRID:SCR_015530) [12] to generate the bam file. The protein and bam files were used as input to Braker for genome annotation.

Platalea minor resequencing and single nucleotide polymorphism analysis

Genomic DNA from 13 P. minor individuals were isolated using the PureLink™ Genomic DNA Mini Kit (Invitrogen Cat no. K182002), following the manufacturer’s instructions. The quality of DNA samples was assessed with the NanoDrop™ One/OneC Microvolume UV–Vis Spectrophotometer and 1% gel electrophoresis. Next, the samples were sent to Novogene for sequencing on an Illumina HiSeq-PE150 platform at approximately 6× coverage. The sequenced raw reads were then trimmed by Trimmomatic (v0.39, RRID:SCR_011848) [13] and cleaned with Kraken 2 (RRID:SCR_005484) [14]. The cleaned reads were aligned to large scaffolds (>500 kb, n = 234), accounting for 97.1% of the P. minor reference genome, with BWA-MEM (RRID:SCR_022192) [15] using the parameters “-t 30 -M -R”. Variant calling was performed using “HaplotypeCaller” and “GenotypeGVCFs” commands from the Genome Analysis Toolkit (GATK, RRID:SCR_001876, v4.1.2.0) [16]. Hard filtering was employed to filter out single nucleotide polymorphisms (SNPs) with the following criteria: quality by depth <2.0, Fisher strand bias >60.0, mapping quality <40.0, mapping quality rank sum test <−12.5, and read position rank sum test <−8.0. The remaining SNPs were further filtered for bi-allelic (“--min-alleles 2 --max-alleles 2”), no missing data (“--max-missing 1”), minimum summed site-depth (sumDP) of 20 and maximum sumDP of 130 to remove sites that were below one-third and above three-fold of the average sumDP, respectively, using the “--site-depth” and “--positions” options in VCFtools (v0.1.16, RRID:SCR_001235) [17]. The heterozygosity and inbreeding coefficient were estimated using VCFtools (v0.1.16) [17]. Details of the resequencing data are listed in Table 1.

Data validation and quality control

During DNA extraction and PacBio library preparation, the samples were subjected to quality control with NanoDrop™ One/OneC Microvolume UV–Vis Spectrophotometer, Qubit^® Fluorometer, and overnight pulse-field gel electrophoresis. The Omni-C library was inspected by Qubit^® Fluorometer and TapeStation D5000 HS ScreenTape.

Regarding the genome assembly, the Hifiasm output was blast to the NT database, and the resulting output was used as input for Blobtools (v1.1.1, RRID:SCR_017618) [18]. Scaffolds identified as possible contaminations were removed from the assembly manually (Figure 2). A statistical kmer-based approach was applied to estimate the heterozygosity of the assembled genome. The repeat content and the corresponding sizes were analysed with k-mer 21 using Jellyfish (RRID:SCR_005491) [19] and GenomeScope (RRID:SCR_017014) [20] (Figure 3; Table 2). BUSCO (v5.5.0) [21] was used to assess the completeness of the genome assembly and gene annotation with a metazoan dataset (aves_odb10). HiC contact maps were generated using Juicer tools (version 1.22.01, RRID:SCR_017226) [22], following the Omni-C manual [23].

Figure 2.

Genome assembly quality control and contaminant/cobiont detection. The upper panel shows the BlobPlot of the assembly. Each circle represents a scaffold with its size scaled according to its scaffold length, while the colour of the circle indicates the taxonomic assignment from BLAST similarity search results. The lower panel reveals the ReadCovPlot of the assembly, illustrating the proportion of unmapped and mapped sequences in the BLAST similarity search results on the left. The latter is further dissected according to the rank of phylum on the right.

Figure 3.

The GenomeScope profile with kmer 21.

Table 2.

Summary of the GenomeScope statistics (k = 21).

Property	min	max
Homozygous (aa)	99.34%	99.37%
Heterozygous (ab)	0.63%	0.66%
Genome Haploid Length (bp)	1,141,324,739	1,144,280,536
Genome Repeat Length (bp)	112,880,770	113,173,108
Genome Unique Length (bp)	1,028,443,969	1,031,107,427
Model Fit	93.18%	99.52%
Read Error Rate	0.41%	0.41%

Omni-C reads and PacBio HiFi reads were used to measure the assembly completeness and the consensus quality (QV) using Merqury (v1.3, RRID:SCR_022964) [24] with kmer 20, resulting in 95.0738% kmer completeness for the Omni-C data and 59.746 QV scores for the HiFi reads, corresponding to 99.999% accuracy.

The black-faced spoonbill genome assembly was also compared to five other avian genomes with chromosome-level assemblies and genome annotations, including Gallus gallus (Ggal: GCF_016699485.2), Cuculus canorus (Ccan: GCF_017976375.1), Mycteria americana (Mame: GCA_035582795.1), Taeniopygia guttata (Tgut: GCF_003957565.2), and Theristicus caerulescens (Tcae: GCA_020745775.1), which were downloaded from NCBI [25] and UCSC [26], respectively [27–29]. Macrosynteny was performed using MCScan (RRID:SCR_017650) with default parameters [30]. It is worth noting that some parts of the largest scaffold in P. minor mapped to several T. caerulescens chromosomes, while genomes in other birds show relatively high syntenic conservation [31], which may warrant further investigation (Figure 4).

Figure 4.

Macrosynteny between P. minor (Pmin), Theristicus caerulescens (Tcae), Mycteria americana (Mame), Taeniopygia guttata (Tgut), Cuculus canorus (Ccan), and Gallus gallus (Ggal).

Results and discussion

Genome assembly of P. minor

A total of 25.35 Gb of HiFi bases was generated with an average HiFi read length of 9,365 bp with 20× data coverage (Table 1). After scaffolding with 77.79 Gb Omni-C sequencing data, the assembled genome size was 1.24 Gb in 468 scaffolds, with a scaffold N50 of 53 Mb and L50 of 8 (Tables 1, 3 and 4; Figures 1B and C). The genome size is comparable to those of other bird species in the family Threskiornithidae, which have genome sizes around 1.0–1.3 Gb, according to the data available in the NCBI Genbank, such as Theristicus caerulescens (1.20 Gb, GCA_020745775.1), Nipponia nippon (1.31 Gb, GCA_035839065.1), and Mesembrinibis cayennensis (1.19 Gb, GCA_013399675.1). The genome completeness was estimated by BUSCO (RRID:SCR_015008) with a value of 97.3% (aves_odb10) (Table 3; Figure 1B). The GC content was 42.98%. A total of 14,673 gene models were generated with 18,780 predicted protein-coding genes, having a mean coding-sequence length of 516 amino acids and a complete protein BUSCO value of 78.3% (Table 3).

Table 3.

Genome statistics.

	Platalea minor
Total length (bp)	1,239,504,613
Number	468
Mean length (bp)	2,648,514
Longest scaffold length (bp)	108,170,464
Shortest scaffold length (bp)	1,000
N_count	0.02%
Gaps	998
N50	53,081,851
N50n	8
N70	31,851,659
N70n	15
N90	14,707,644
N90n	25
BUSCO (Geno, metazoa_odb10)	C:93.7%[S:93.4%,D:0.3%],F:1.6%,M:4.7%,n:954
BUSCO (Geno, aves_odb10)	C:97.3%[S:97.0%,D:0.3%],F:0.5%,M:2.2%,n:8338
Protein total length (amino acids)	9,684,019
Protein number (amino acids)	18,780
Protein mean length (amino acids)	516
BUSCO (Prot, metazoa_odb10)	C:88.4%[S:72.5%,D:15.9%],F:1.7%,M:9.9%,n:954
BUSCO (Prot, aves_odb10)	C:78.3%[S:59.9%,D:18.4%],F:1.9%,M:19.8%,n:8338

Table 4.

Scaffold information with a length larger than 1 Mb.

Scaffold ID	Scaffold length (bp)	Cumulative % of the whole genome
scaffold_1	108,170,464	8.73%
scaffold_2	94,402,276	16.34%
scaffold_3	92,629,600	23.81%
scaffold_4	79,211,724	30.20%
scaffold_5	73,048,200	36.10%
scaffold_6	72,537,701	41.95%
scaffold_7	69,784,815	47.58%
scaffold_8	53,081,851	51.86%
scaffold_9	44,005,081	55.41%
scaffold_10	36,492,200	58.35%
scaffold_11	35,351,409	61.20%
scaffold_12	34,190,505	63.96%
scaffold_13	33,557,343	66.67%
scaffold_14	33,263,400	69.35%
scaffold_15	31,851,659	71.92%
scaffold_16	31,385,420	74.45%
scaffold_17	30,450,346	76.91%
scaffold_18	30,280,111	79.35%
scaffold_19	26,443,200	81.49%
scaffold_20	25,001,000	83.50%
scaffold_21	23,104,878	85.37%
scaffold_22	18,730,967	86.88%
scaffold_23	17,395,620	88.28%
scaffold_24	15,031,567	89.49%
scaffold_25	14,707,644	90.68%
scaffold_26	14,565,521	91.86%
scaffold_27	13,184,709	92.92%
scaffold_28	9,343,000	93.67%
scaffold_29	7,143,856	94.25%
scaffold_30	6,869,713	94.80%
scaffold_31	6,476,723	95.33%
scaffold_32	4,846,626	95.72%
scaffold_33	4,355,004	96.07%
scaffold_34	2,327,616	96.26%
scaffold_35	2,194,594	96.43%
scaffold_36	1,422,971	96.55%
scaffold_37	1,263,285	96.65%
scaffold_38	1,176,292	96.74%
scaffold_39	1,115,873	96.83%

Repeat content

A total repeat content of 11.94% was found in the genome, which contained a lower level of repeat elements, similar to other avian genomes [32], with 2.49% unclassified elements. Of the remaining repeats, long interspersed nuclear elements (LINE) were the most abundant (5.10%), followed by long terminal repeats (LTR) (1.62%). In contrast, DNA, short interspersed nuclear elements (SINE), Penelope, and rolling circle were only present in low proportions (DNA: 0.63%, SINE: 0.09%, Penelope: 0.06%, rolling circle: 0.02%). A complete catalogue of the repeat content of the genome can be found in Table 5 and Figure 1D.

Table 5.

Summary of the repetitive elements analysis.

Classification	Coverage length (bp)	Count	Proportion (%)	No. of distinct classifications
DNA	7,812,155	25,610	0.63	3,780
LINE	63,189,657	102,036	5.10	5,652
LTR	20,129,006	29,922	1.62	3,928
Other (Simple Repeat, Microsatellite, RNA)	23,864,680	3,875	1.93	1,118
Penelope	690,955	1,966	0.06	646
Rolling Circle	287,526	744	0.02	404
SINE	1,165,920	4,910	0.09	885
Unclassified	30,908,397	60,141	2.49	5,695
SUM	148,048,296	229,204	11.94	22,108

Single nucleotide polymorphism sites

A total of 6,046,878 bi-allelic SNPs were called from 13 P. minor individuals, accounting for ∼0.5% of the genome. The mean individual heterozygosity was 0.142%. The lowest individual heterozygosity (0.077%) was close to other endangered bird species, such as Pelecanus crispus (0.60%) and Nestor notabilis (0.91%) [33]. The heterozygosity levels (0.108% to 0.116%) from five individuals were comparable to previous reports on spoonbills - black-faced spoonbill (0.101%–0.116%, mean 1.09%, n = 11) and royal spoonbill (0.098%–0.109%, mean 0.105%, n = 9) [4]. The remaining heterozygosity levels observed in this study were below the mean (0.221%) and median (0.213%) of heterozygosity reported from 40 avian species [33]. Signals of inbreeding were observed among the samples, with the inbreeding coefficient (F _IS) ranging from 0.331 to 0.720 (Table 6), providing additional evidence of a recent genetic bottleneck in the black-faced spoonbill population [4]. High levels of F _IS have also been observed in other bird populations suffering from past bottlenecks [34]. These results highlighted the need for continuous efforts in monitoring P. minor.

Table 6.

Number of SNPs, statistics of heterozygosity and inbreeding coefficient of 13 Platalea minor individuals.

Sample ID	No. of sites with observed heterozygosity	H 0 (%)	F IS
BFS1	1,015,586	0.168	0.382
BFS2	673,637	0.111	0.590
BFS3	654,457	0.108	0.602
BFS4	1,013,752	0.168	0.383
BFS5	700,877	0.116	0.573
BFS6	1,076,231	0.178	0.345
BFS7	1,080,566	0.179	0.342
BFS8	1,055,650	0.175	0.357
BFS9	463,725	0.077	0.718
BFS10	997,870	0.165	0.393
BFS11	700,069	0.116	0.574
BFS12	652,122	0.108	0.603
BFS14	1,109,354	0.183	0.325

Conclusion and reuse potential

This study presents the first chromosomal-level genome assembly and single-nucleotide polymorphism sites of black-faced spoonbill Platalea minor. These are useful and valuable resources for future population genomic studies aimed at better understanding spoonbill species numbers and conservation.

Funding Statement

This work was funded and supported by the Hong Kong Research Grant Council Collaborative Research Fund (C4015-20EF), CUHK Strategic Seed Funding for Collaborative Research Scheme (3133356) and CUHK Group Research Scheme (3110154).

Data availability

The final assembly has been deposited at NCBI under the accession number JBBPFK000000000. The raw reads generated in this study, including Omni-C (SAMN40731791) and PacBio HiFi (SAMN35152374) data, have been deposited in the NCBI database under the BioProject accession number PRJNA973839. The genome, genomic and repeat annotation files have been deposited and are publicly available in Figshare [35].

Abbreviations

HiFi, high-fidelity; HMW, high molecular weight; LINE, long interspersed nuclear element; LTR, long terminal repeat; QV, consensus quality; SINE, short interspersed nuclear element; SNP, single nucleotide polymorphisms; sumDP, summed site-depth; TE, transposable elements.

Declarations

Ethics approval and consent to participate

The authors declare that ethical approval was not required for this type of research.

Competing interests

The authors declare that they do not have competing interests.

Authors’ contributions

JHLH, TFC, LLC, SGC, CCC, JKHF, JDG, SCKL, YHS, CKCW, KYLY and YW conceived and supervised the study; WLS carried out DNA extraction, library preparation and sequencing; WN performed genome assembly and gene model prediction; STSL carried out the SNPs calling and Fst calculations; PC, AL, LRJ and HYY collected and maintained the samples. All authors approved the final version of the manuscript.

Funding

This work was funded and supported by the Hong Kong Research Grant Council Collaborative Research Fund (C4015-20EF), CUHK Strategic Seed Funding for Collaborative Research Scheme (3133356) and CUHK Group Research Scheme (3110154).