Comparative Genome and Evolution Analyses of an Endangered Stony Coral Species Dendrophyllia cribrosa Near Dokdo Islands in the East Sea

Abstract Stony corals often harbor intracellular photosynthetic dinoflagellate algae that receive dissolved inorganic nutrients. However, Dendrophyllia cribrosa is a nonsymbiotic stony coral distributed in the western Pacific. We assembled a chromosome-level D. cribrosa genome using PacBio and Hi-C technologies. The final assembly was 625 Mb, distributed on 14 chromosomes, and contained 30,493 protein-coding genes. The Benchmarking Universal Single-Copy Orthologs analysis revealed a percentage of 96.8 of the metazoan genome. A comparative phylogenetic analysis revealed that D. cribrosa, which lacks symbionts, evolved to acquire cellular energy by expanding genes related to acyl-CoA metabolism and carbohydrate transporters. This species also has expanded immune-related genes involved in the receptor protein tyrosine kinase signaling pathway. In addition, we observed a specific expansion of calcification genes, such as coral acid-rich proteins and carbonic anhydrase, in D. cribrosa. This high-quality reference genome and comparative analysis provides insights into the ecology and evolution of nonsymbiotic stony corals.


Introduction
Dendrophyllia cribrosa, belonging to the scleractinian coral family, is a rare subtropical-temperate coral species that is distributed in the western Pacific. Dendrophyllia cribrosa is a stony coral without symbiotic microalgae. In 2016, the Ministry of Oceans and Fisheries of Korea reported a single habitat of a D. cribrosa coral community with a width of 5 m and a height of 3 m, at depths of 18-20 m, near the Dokdo Islands in the East Sea. The morphological features of D. cribrosa resemble trees with irregular thick branches ( fig. 1a), and their coloration ranges from deep yellow to orange. This species was designated as endangered in the "Endangered and Protected Wild Species List in Korea" in 1998 by the Korean Government.
Here, we describe a chromosome-level assembly of D. cribrosa from the Dokdo Islands in Korea. The Dendrophyllia cribrosa genome provides a comparative study of coral genomes that exhibit evolutionary expansions related to coral calcification, metabolism and immune responses.

Genome Assembly of Dendrophyllia cribrosa
We produced 41 Gb next-generation sequencing (NGS) reads of D. cribrosa (supplementary table S1, Supplementary Material online). To estimate the genome size of D. cribrosa, we used Jellyfish, programmed with a K-mer range of 17-25. Jellyfish estimated the genome size of D. cribrosa to be 610 Mb, at K = 25, with the lowest PCR error rate (0.19) and PCR duplicates (0.92), which is similar to the genome size of other closely related complex corals (Montipora spp. 615-653). We added the GenomeScope result in supplementary figure S1, Supplementary Material online. At K = 25, we estimated the D. cribrosa genome size to be 610 Mb with 0.30% heterozygosity in 25 bp of K-mer (supplementary fig. S1, Supplementary Material online). This estimation is similar to the genome size of closely related complex corals (Genus Montipora, 615-653 Mb) (Helmkampf et al. 2019). We also produced 120 Gb-long reads (∼246-fold coverage of the genome) using a PacBio Sequel2 platform (DNA Link Inc., Seoul, Republic of Korea) with an N50 of 27 kb (supplementary  table  S2, Supplementary Material online). The FALCON_unzip assembler constructed 1,174 contigs, with an assembly length of 765 Mb (table 1). After implementing purge haplotigs and error correction, we obtained a 680M assembly from 591 contigs. The Benchmarking Universal Single-Copy Orthologs (BUSCO) assessments showed that the number of "complete and duplicated BUSCO genes" was slightly decreased from 23 (9.0%) to 10 (3.9%) without any change in the total number of complete genes. After polishing the haplotigs, we could not find any changes in BUSCO values (table 1). The N50 of our contigs was 2.

GBE
To resolve this, we manually split these pseudo-scaffolds (supplementary fig. S2b, Supplementary Material online) and removed seven contigs to generate the separated scaffolds. Additionally, we did not use any contigs that were <1 kb in the BUSCO assessment, which resulted in a lower score (93.7%). The N50 of the D. cribrosa assembly was 19 Mb, and the maximum assembly length was 62 Mb. Based on the BUSCO assessment score, we measured 93.7% completeness of genes, including 92.5% completeness of 236 singlecopy genes and 1.2% completeness of three duplicated BUSCO genes. We found 14 (0.8%) missing genes in 14 pseudo-chromosomes. During the scaffolding, several genes were not integrated in the scaffolds.
Approximately 364 Mb (58.10%) of repeats were found in the D. cribrosa genome (supplementary table S3, Supplementary Material online). This proportion is similar to that of other coral genomes, such as Trachythela (57.88%) (Zhou et al. 2021). We predicted 30,493 proteincoding genes in the D. cribrosa genome from these data. They showed a slightly higher number of protein-coding genes compared with other coral genes (supplementary table S4, Supplementary Material online). We conducted BUSCO assessment in the protein-coding genes. It resulted in 231 (90.6%) complete eukaryote genes, with 226 (88.6%) single copy and 5 (2.0%) duplicated in the BUSCO gene set. Among them, 8 (3.1%) were fragmented and 16 (6.3%) were missed. The BUSCO assessment showed a higher number of complete genes in the D. cribrosa genome (supplementary table S4, Supplementary Material online). A long-read sequence library was constructed using the SMRTbell Express Template Preparation Kit (101-357-000) and sequenced using the PacBio Sequel2 platform. An Arima-Hi-C kit (Arima Genomics Inc., San Diego, CA, USA) was used according to the manufacturer's instructions. The Hi-C library was sequenced using the NovaSeq 6000 platform (Novogene Co. Ltd, CA, USA).

Supplementary Material
Supplementary data are available at Genome Biology and Evolution online.