2 | Cane Toad Genome
The final hybrid assembly of 31,392 scaffolds was 2.55 Gb in length with a scaffold N50 of 168 kb. BUSCO analysis revealed that the assembly included full length or partial fragments of 90.6% of tetrapod universal single-copy orthologs (n = 3950), illustrating that the gene-containing regions have been well assembled. Annotation predicted 25,846 protein coding genes with similarity to known proteins in Swiss-Prot. Repeat sequences were estimated to account for 63.9% of the assembly. Conclusions: The R. marina draft genome assembly will be an invaluable resource that can be used to further probe the biology of this invasive species. Future analysis of the genome will provide insights into cane toad evolution and enrich our understanding of their interplay with the ecosystem at large.
Keywords: cane toad; Rhinella marina; sequencing; hybrid assembly; genome; annotation
Data Description
Introduction
The cane toad (Rhinella marina formerly Bufo marinus) (Fig. 1) is a true toad (Bufonidae) native to Central and South America that has been introduced to many areas across the globe [1]. Since its introduction into Queensland in 1935, the cane toad has spread widely and now occupies more than 1.2 million square kilometers of the Australian continent, fatally poisoning predators such as the northern quoll, freshwater crocodiles, and several species of native lizards and snakes [1-5]. The ability of cane toads to kill predators with toxic secretions has contributed to the success of their invasion [1]. To date, research on cane toads has focused primarily on ecological impacts, rapid evolution of phenotypic traits, and population genetics using neutral markers [6, 7], with limited knowledge of the genetic changes that allow the cane toad to thrive in the Australian environment [8-11]. A reference genome will be useful for studying loci subject to rapid evolution and could provide valuable insights into how invasive species adapt to new environments. Amphibian genomes have a preponderance of repetitive DNA [12, 13], confounding assembly with the limited read lengths of first- and second-generation sequencing technologies. Here, we employ a hybrid assembly of Pacific Biosciences (PacBio) long reads and Illumina short reads (Fig. (Fig.2)2) to overcome assembly challenges presented by the repetitive nature of the cane toad genome. Using this approach, we assembled a draft genome of R. marina that is comparable in contiguity and completeness to other published anuran genomes [14-17]. We used our previously published transcriptomic data [18] and other published anuran sequences to annotate the genome. Our draft cane toad assembly will serve as a reference for genetic and evolutionary studies and provides a template for continued refinement with additional sequencing efforts.
Sample collection, library construction, and sequencing
Adult female cane toads were collected by hand from Forrest River in Oombulgurri, WA (15.1818oS, 127.8413oE) in June 2015. Toads were placed in individual damp cloth bags and transported by plane to Sydney, NSW, before they were anaesthetized by refrigeration for 4 hours and killed by subsequent freezing. High-molecular-weight genomic DNA (gDNA) was extracted from the liver of a single female using the genomic-tip 100/G kit (Qiagen, Hilden, Germany). This was performed with supplemental RNase (Astral Scientific, Taren Point, Australia) and proteinase K (NEB, Ipswich, MA, USA) treatment, as per the manufacturer's instructions. Isolated gDNA was further purified using AMPure XP beads (Beckman Coulter, Brea, CA, USA) to eliminate sequencing inhibitors. DNA quantity was assessed using the Quanti-iT PicoGreen dsDNA kit (Thermo Fisher Scientific, Waltham, MA, USA). DNA purity was calculated using a Nanodrop spectrophotometer (Thermo Fisher Scientific), and molecular integrity was assessed using pulse-field gel electrophoresis.
For short-read sequencing, a paired-end library was constructed from the gDNA using the TruSeq polymerase chain reaction (PCR)-free library preparation kit (Illumina, San Diego, CA, USA). Insert sizes ranged from 200 to 800 bp. This library was sequenced (2 × 150 bp) on the HiSeq X Ten platform (Illumina) to generate approximately 282.9 Gb of raw data (Table (Table1).1). Illumina short sequencing reads were assessed for quality using FastQC v0.10.1 [19]. Low-quality reads were filtered and trimmed using Trimmomatic v0.36 [20] with a Q30 threshold (LEADING:30, TRAILING:30, SLIDINGWINDOW:4:30) and a minimum 100-bp read length, leaving 64.9% of the reads generated, of which 75.2% were in retained read pairs.
For long-read sequencing, we used the single-molecule real-time (SMRT) sequencing technology (PacBio, Menlo Park, CA, USA). Four SMRTbell libraries were prepared from gDNA using the SMRTBell template preparation kit 1.0 (PacBio). To increase subread length, either 15–50 kb or 20–50 kb BluePippin size selection (Sage Science, Beverly, MA, USA) was performed on each library. Recovered fragments were sequenced using P6C4 sequencing chemistry on the RS II platform (240 min movie time). The four SMRTbell libraries were sequenced on 97 SMRT cells to generate 7,745,233 subreads for 76.6 Gb of raw data. Collectively, short- and long-read sequencing produced around 359.5 Gb of data (Table 1).
Genome assembly
We employed a hybrid de novo whole-genome assembly strategy, combining both short-read and long-read data. Trimmed Q30-filtered short reads were de novo assembled with ABySS v1.3.6 [21] using k = 64 and default parameters (contig N50 = 583 bp) (Table 2). Long sequence reads were de novo assembled using the program DBG2OLC [22] (k 17 AdaptiveTh 0.0001 KmerCovTh 2 MinOverlap 20 RemoveChimera 1) (contig N50 = 167.04 kbp) (Ta-
