Genome sequence of the Mediterranean red coral Corallium rubrum

Abstract

Objectives

Corallium rubrum, the precious red coral, is an octocoral endemic to the western Mediterranean Sea. Like most octocorals, it produces tiny, calcified structures called sclerites. Uniquely, it also produces a completely calcified axial skeleton that is a bright red color. This combination of color and hardness has made the red coral prized for centuries, leading to extensive fishing and trade for use in jewelry. Understanding how it produces this red skeleton is thus a central question in economics, culture, and biology. To gain insights into this process, we sequenced the C. rubrum genome.

Data description

Our C. rubrum genome assembly is 655 megabases (Mb) in size, distributed across 2910 scaffolds with a very low level of unknown nucleotides (0.95%). We used a pipeline based on the MaSuRCA hybrid assembler, combining long PacBio reads and short Illumina reads, followed by several steps to improve the assembly, including scaffolding, merging, and polishing. This represents the third published genome of an octocoral and the first within the order Scleralcyonacea.

Objective

The precious red coral, Corallium rubrum, holds a special place in history. It was the first "coral" to be identified, lending its name to all other corals. Its use dates back to Neolithic times, valued not only for its beauty in jewelry but also for its symbolic, cultural, and even medicinal significance [1, 2].

C. rubrum plays a key role in the well-known Mediterranean coralligenous ecosystem, a biodiversity hotspot. This slow-growing species (millimeters per year) thrives at various depths, ranging from 5 to 800 m, with populations structured according to depth zone [3, 4]. However, overfishing and global warming have led to the decrease of its population, prompting the IUCN to classify it as Endangered [5].

Red corals (family Coralliidae) possess a unique trait: they produce sclerites (tiny calcified elements) like other octocorals, but also a fully calcified axial skeleton [6]. This skeleton’s color can range from white (e.g., Corallium konojoi) to bright red, with C. rubrum being particularly prized for its intense hue [7, 8]. However, the genetic mechanisms controlling both calcification and color remain to be elucidated.

Phylogenetically, the subphylum Anthozoa (phylum Cnidaria) encompasses two classes: Hexacorallia (sea anemones, reef-building corals) and Octocorallia (over 3500 species including soft corals, gorgonians, sea pens) [9]. The Hexacorallia phylogeny is now well resolved thanks to extensive genetic sequencing [10,11,12,13,14,15]. Conversely, despite recent revisionary systematics of Octocorallia using targeted capture of specific genetic sequences [16], only two octocorallian genomes [17, 18] (and a few transcriptomes) have been published to date.

Understanding the C. rubrum genome is crucial for further investigations into its biology (e.g., skeleton formation and color, genetic diversity), ecology, and evolutionary history. This knowledge is essential for developing effective conservation strategies to safeguard these vital marine organisms in the face of ongoing environmental challenges [19].

Data description

We report here the complete genome sequence of the red coral, Corallium rubrum, consisting of 655.3 megabases (Mb) assembled into 2910 scaffolds [20] (Table 1). The genomic DNA and total RNA were isolated from a single colony collected near Banyuls-sur-Mer, France [21].

Table 1 Overview of data_files/data_sets

Building the genome assembly

The genomic DNA was sequenced with 2 technologies: Illumina short reads (Paired-End and Mate-Pair) and PacBio long reads [27]. To construct the genome assembly, we initially attempted separate assemblies using only PacBio long reads (Canu-1.6 assembler [29]) or only Illumina short reads (MaSuRCA-3.2.2 assembler [30]) [22]. These initial assemblies had low N50 values (approximatively 36 kb and 21 kb, respectively), indicating a high degree of fragmentation.

To overcome this limitation, we employed a hybrid assembly approach using both PacBio long reads and Illumina short reads with the MaSuRCA hybrid assembler (genome version 1.5.2). This approach yielded a significantly improved assembly with a larger size (686.6 Mb) distributed across 9,364 scaffolds. Importantly, the percentage of unknown nucleotides (“Ns”) significantly decreased from 37.6% to 1.56% between the Illumina-only and the hybrid mode of MaSuRCA [22,23,24,25].

Scaffolding, merging, and polishing

Following the initial assembly, we employed a three-phase process to further refine and improve the genome sequence. The first phase involved scaffolding, which utilized information from mRNA sequences, RNA libraries, and a genome mate-pair library (using L_RNA scaffolder [31], Rascaf [32], and Sspace [33], respectively). This process successfully reduced the number of scaffolds by nearly half [23, 26].

In the second phase, we merged the V1.5.2 scaffolded assembly with the previous Illumina-only and PacBio-only assemblies using Metassembler, an algorithm that merges multiple genome assemblies into a single superior sequence [34].

The final phase focused on polishing the assembly using Pilon [35] and GapCloser [36]. These tools filled remaining gaps and polished the sequence, resulting in the final high-quality genome sequence (less than 1% Ns) of Corallium rubrum.

Data associated with each step of the pipeline is available in Data_Files2–5 [22,23,24,25,26].

Conclusions

A 2022 study revised the octocorallian phylogeny, establishing two orders: Scleralcyonacea and Malacalcyonacea [16]. However, only two complete octocorallian genomes have been published to date, both belonging to the order Malacalcyonacea: Dendronephthya gigantea (286 Mb) [18] and Xenia sp. (222 Mb) [17]. We present the first published genome sequence of a member of the order Scleralcyonacea, Corallium rubrum. The C. rubrum genome is estimated to be 2–3 times larger than previously sequenced octocorallian genomes (but similar in size to other scleractinian corals). Remarkably, transcriptome data suggests that it encodes for around 30,000 protein-coding genes [37, 38], a similar number for all three species. This indicates that the larger genome size of C. rubrum is likely attributed to an abundance of non-coding DNA, especially repetitive elements. These repetitive elements can significantly hinder De Bruijn graph-based genome assemblers reliant on short Illumina reads for assembly [39]. In support of this, we employed various assemblers and only achieved satisfactory metrics when we used MaSuRCA and a combination of short and long reads.

Limitations

The non-annotated Corallium rubrum genome

Our current Corallium rubrum genome sequence data has a key limitation: lack of predicted and annotated genes. Transcribed genes need to be identified downstream using either prediction algorithm and/or transcriptomic data. The first C. rubrum transcriptome was published in 2015 with another proposed in 2016 [37, 38]. In addition, to facilitate gene identification, we provide a link for the alignment (STAR [40]) of RNA reads against the C. rubrum genome (dataset 2, [28]), from which expressed genes can be extrapolated.

Haploid genome

Coral genomes are diploid and often highly heterozygous, in addition to sometimes display high level of tandemly duplicated genes and repeated sequences [13, 15, 41]. While we present a haploid representation of the genome, some scaffold portions might contain biased sequences resulting from uncaptured allelic duplications or misassembled repeats. Unfortunately, our current sequencing and assembling technology could not resolve a fully diploid assembly.

Forthcoming improvements

While advancements in sequencing technology, particularly long-read sequencing and assembling, promise improved accuracy and capability to handle repetitive regions [42], we are confident that our current C. rubrum genome assembly remains an accurate and valuable resource for various research areas, including genomics, metagenomics, and phylogenetic studies.