Objective

Brasenia schreberi is an aquatic and perennial herb in the Cabombaceae family. It is a monotypic species with oval-shaped leaves that can submerge or float on the water’s surface, similar to water lilies. Except for Europe and Antarctica, it is currently distributed on all continents of the world [1]. However, palaeobotanical records indicate that B. schreberi was a frequent element in Europe before the last glacial period [1]. Its habitats include ponds, lakes, and sluggish streams, but they must be clean and acidic and have nutrient-enriched sediment [1, 2]. Due to the deterioration of water quality and habitat loss, it is listed at the second level of national key protected wild plants in China and is endangered in other countries [2, 3]. Its edible young leaves and stems are coated with a thick mucilage that is mainly composed of polysaccharides and has high medical value [2, 4, 5]. Mucilage has been confirmed to be a defense against herbivores and bacteria [3, 6]. Brasenia schreberi contains allelopathic components that can be used in the management of aquatic weeds [7]. As important values in taxonomy, ecology, and economy and in its endangered situation, a genome assembly was published previously [8] for its better conservation and breeding. However, given its wide distribution worldwide and existing substantial genetic diversities [3, 9], we present an alternative B. schreberi genome to better understand its evolution and adaptation and to enhance its conservation, management, and utility in the future.

Data description

Leaf samples of B. schreberi were collected from an individual planted in the South China Botanical Garden, Guangzhou, China. The DNA or RNA extracted from its leaf tissues was used to construct three sequencing libraries, including long read whole genome sequencing (WGS) using a Nanopore PromethION sequencer, short read WGS using an MGI DNBSEQ-T7 sequencer, and RNA sequencing (RNA-seq) using an MGI DNBSEQ-T7 sequencer. Under the MGI platforms, a 150 bp paired-end mode was applied for both short WGS and RNA-seq. The long-read WGS generated about 113.0 GB of data (Data file 1) [10], short-read WGS generated about 130.6 GB data (Data file 2) [11], and RNA-seq generated about 27.6 GB data (Data file 3) [12].

After sequencing, short WGS reads were trimmed by Sickle v1.33 [13] using the parameter “-q 30 -l 80”. KmerGenie v1.7044 [14] (under the parameter of “-k 141 --diploid”) was then used to estimate the genome size of B. schreberi with trimmed short WGS reads. The estimated genome size was 963,304,542 bp. Porchop v0.2.4 [15] and ontbc v1.1 [16] were used to remove adapter and low-quality reads (scores < 7 and lengths < 5000 bp) in long WGS reads. NextDenovo v2.3.1 [17] was then used to assemble the genome with the filtered long reads. Pseudohaploid [18] and Purge_Dups v1.2.6 [19] were applied to remove redundant contigs. Subsequently, Racon v1.5.0 [20], hapo-G v1.3.2 [21], and polypolish v0.5.0 [22] were used to polish the assembly. The final assembly was 1,055,148,839 bp in length, with 92 contigs and a contig N50 of 22,379,495 bp (Data file 4) [23]. BUSCO v5.5.0 [24] and compleasm v0.2.5 [25] were used to assess the completeness of the assembly with Eudicots odb10-2020-09-10 and Embryphyta odb10 2020-09-10 databases. BUSCO revealed 88.4 and 95.4% completeness in the Eudicots and Embryphyta databases, respectively (Data files 5–6) [26, 27]. Compleasm revealed 90.9 and 96.7% completeness in the Eudicots and Embryphyta databases, respectively (Data files 7–8) [28, 29].

Repetitive elements in the B. schreberi assembly were estimated by RED v2.0 [30] and EDTA v2.1.3 [31], which revealed 452,408,938 (Data file 8) [32] and 521,424,853 bp (Data file 9) [33] of sequences, respectively. Combining the RED and EDTA results revealed 555,442,205 bp of repetitive sequences (Data file 10) [34], which were used to soft-mask the assembly. Braker3 v.3.0.6 [35] was used to predict the primary gene structures using transcriptome data and reference protein sequences (Data file 11) [36]. The Braker results were then incorporated into the Funannotate pipeline v1.8.16 [37] to obtain integrated gene sets. The pipeline included four steps: “train”, “predict”, “update”, and “annotate”. For the former three steps, the parameter “--max_intronlen 1000000” was used, while in the “predict” step, the parameters “--busco_seed_species arabidopsis --organism other --busco_db embryophyta” were added. The fourth “annotate” step was used for gene function annotation. The final gene prediction obtained 67,747 protein-coding genes and 813 tRNA genes (Data files 12–14) [38,39,40]. Functional annotation of protein-coding genes is shown in Data files 15–16 [41, 42].

Limitations

The current B. schreberi assembly in this study is fragmented. Future sequencing technologies, including Hi-C, Nanopore ultra-long sequencing, PacBio HiFi, 10X Genomics linked sequencing, and Bionano optical maps, are needed for complete and gapless genome assembly.

However, our assembly displayed a completeness comparable to the previously reported B. schreberi assembly [8], which showed 89.0 and 95.9% completeness using BUSCO in the Eudicots and Embryphyta databases, respectively, and 91.3% and 97.0% completeness using compleasm in the Eudicots and Embryphyta databases, respectively. Nevertheless, because this previous assembly did not remove duplications from the assembly [43], some assembly errors may exist for gene prediction. For the completed BUSCOs, our assembly revealed 39.7% and 46.8% higher complete and single-copy BUSCOs using BUSCO in the Eudicots and Embryphyta database, while it was 37.4 and 44.0% complete and single-copy BUSCOs in Eudicots and Embryphyta for the previously reported assembly. Using compleasm, our assembly was shown to have 47.9 and 54.7% complete and single-copy BUSCOs in the Eudicots and Embryphyta database, while it was 44.54 and 50.9% complete and single-copy BUSCOs in the Eudicots and Embryphyta database for the previously reported assembly. Therefore, our assembly contained a few duplication errors in the assembly for better gene prediction.

Table 1 Overview of all data files/data sets
Full size table