Illumina sequencing, de novo assembly, and gene annotation

Illumina sequencing of O. margaretae yielded a total of 62,321,166 raw reads. Among them, 3,504,433 were first filtered out before assembly as low quality sequences or potential contaminations. Thirty combinations of multiple K-mer lengths and coverage cut-off values were used to perform de novo assembly of the clean reads [22,23]. Finally, we merged the 30 raw assemblies by integrating sequence overlaps and eliminating redundancies. The final assembly included a total of 54.3 mega base pairs (Mb), and 37,906 transcripts were obtained with a N50 length of 1,870 base pairs (bps) and a mean length of 1,434 bps. The sequencing information is presented in Table 1 and the length distribution of all transcripts is shown in Figure 1. All sequence reads are deposited at NCBI (accession number SRA091981), and the final assembly is presented as supporting information (Sequence S1 and S2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Length distribution of transcripts in base pairs.

The numbers of transcripts are shown on top of each bar.

https://doi.org/10.1371/journal.pone.0075211.g001

A reference dataset was constructed for gene annotation, which included protein data from seven species (Anolis carolinensis, Danio rerio, Gallus gallus, Homo sapiens, Mus musculus, Oryzias latipes, and Xenopus tropicalis), representing all major lineages of vertebrates. Transcripts were blasted against this dataset. In total, 18,933 transcripts were annotated to 14,628 genes, which comprised 49.95% of the total transcripts. E-value distribution showed that 73.89% of the annotated sequences had strong homology (E-value below 1E-50), and similarity distribution showed that 72.50% of the annotated sequences had a similarity greater than 60% (Table 1 and Figure 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Characteristics of gene annotation of assembled transcripts against the reference dataset.

(A) E-value distribution of BLASTX hits for transcript with a cut-off E-value of 1E-5. (B) Similarity distribution of BLASTX hits for transcript.

https://doi.org/10.1371/journal.pone.0075211.g002

GO classification

Gene Ontology (GO) is widely used to standardize representation of genes across species and provides a set of structured and controlled vocabularies for annotating genes, gene products, and sequences [24]. In total, 11,457 unique transcripts were assigned to 52 level-2 GO terms, which were summarized under three main GO categories, including cellular component, molecular function, and biological process (Figure 3). Compared to the GO annotations of two other species of the same family, R. chensinensis and R. kukunoris, the GO category distributions of the transcripts for the three ranid frogs were highly similar (Figure S1). Within the GO category of cellular components, 14 level-2 categories were identified, and the terms cell, cell part, and organelle were the most abundant (>50%). Within the GO category of molecular function, 15 level-2 categories were identified, and the term binding was the most abundant (>50%). For biological process function, 23 level-2 categories were identified, and the terms of cellular process and metabolic process were the most abundant (>50%).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Distribution of Gene Ontology (GO) categories (level 2) of transcripts for O. margaretae.

GO functional annotations are summarized in three main categories: cellular component, molecular function and biological process.

https://doi.org/10.1371/journal.pone.0075211.g003

KEGG analysis

Kyoto Encyclopedia of Genes and Genomes (KEGG) [25] database was used to identify potential biological pathways represented in the O. margaretae transcriptome. A total of 1,438 transcripts were assigned to 128 KEGG pathways (Figure 4, Table S1). Among the pathways, purine metabolism, pyrimidine metabolism, phosphatidylinositol signaling system, and a few others were highly represented. These annotations provide a valuable resource for investigating specific processes, functions, and pathways in amphibian research.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Distribution of O. margaretae transcripts among Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways.

The top 16 most highly represented pathways are shown. Analysis was performed using Blast2GO and the KEGG database.

https://doi.org/10.1371/journal.pone.0075211.g004

Comparison to other amphibian transcriptomes

Overall, our data were similar to other reported amphibian transcriptomes from Illumina sequencing platform (Table 2). We had a slightly longer N50 length than most of the other transcriptomes, suggesting the quality of our assembly was high. Clearly, our strategy with multiple K-mer length and cut-off value combinations worked well. The transcriptomes of R. chensinensis [12] and Hyla arborea [14] had short N50 lengths and large numbers of transcripts, suggesting that the assembly generated many short transcripts and could be improved. Among the eight transcriptomes, C. alboguttata had the largest number of transcripts; this is not particularly surprising, considering that it also had the largest number of raw reads. In terms of transcript annotation rate, O. margaretae had approximately 50% transcripts that were annotated, whereas the rates were approximately 33%, 40%, 44%, 42% and 32% for C. alboguttata [10], R. chensinensis [11], R. chensinensis [12], R. kukunoris [11], and N. viridescens [15], respectively.

AMPs

A total of 27 transcripts were identified as potential AMPs (Table S2). Among them, five transcripts matched AMPs that were previously identified in O. margaratae, including Esculentin-2-OMar1, Brevinin-2-OMar1, Odorranain-A-Omar1, Esculentin-l-OMar4, and Margaratain-C1. The other 22 were new to O. margaratae. Although the exact functions of these peptides remain largely unknown, AMPs have received much attention lately [20,26], due to their anti-infective properties.

Molecular markers

A total of 50,351 single nucleotide polymorphism (SNP) sites were identified; 33,735 were transitions and 16,616 were transversions (Table S3). We also identified 2,574 microsatellite DNA loci, of which 78.59% were dinucleotide repeats, 19.54% were trinucleotide repeats, 1.40% were tetranucleotide repeats, 0.43% were pentanucleotide repeats, and 0.04% were hexanucleotide repeats (Table S4). These molecular markers will provide useful tools in population genetic and molecular ecological studies of O. margaretae, and potentially for other odorous frogs as well.

Sample collection

In order to recover as many expressed genes as possible, multiple tissues from three adult individuals (one male and two females) and one tadpole were used. For adults, seven types of tissues (cerebrum, eye, skeletal muscle, heart, liver, testicle, and Ootheca) were collected; for tadpole, the whole body was used after removing the guts. Samples of one male, one female, and the tadpole were collected from Mt. Emei (Sichuan, China; E103°38918’, N29° 56418’, 749m) in August 2012. The other female was collected from Xiaohegou Nature Reserve (Sichuan, China; E104°47257’, N32° 53149’, 1393m; approximately 345 km from the first site) in September 2012. All individuals were euthanized by immersion in MS-222 buffered solution (3g/L), and tissues were collected and stored in Sample Protector Solution (TAKARA) immediately after euthanasia. All individuals were identified by both morphological and molecular (mitochondrial cytochrome b gene sequence) traits.

cDNA library construction and Illumina sequencing

Total RNA was extracted from each tissue sample individually using TRIzol® reagent (Life Technologies) and mixed with approximately same quantity. RNA integrity was assessed using the RNA 6000 Nano Assay Kit with a Bioanalyzer 2100 (Agilent Technologies) after checking the RNA purity and concentration. A single cDNA library was constructed. The mRNA was purified from total RNA using poly-T oligo-attached magnetic beads (Life Technologies). The first cDNA strand was synthesized using random oligonucleotides and M-MuLV Reverse Transcriptase (RNase H-). The second cDNA strand synthesis was subsequently performed using DNA Polymerase I and RNase H. The cDNA with an insert size of 200 bp were preferentially purified with AMPure XP beads system (Beckman Coulter) and sequenced on an Illumina HiSeq 2000 platform. 101 bp paired-end reads were generated and all raw sequence read data were stored in FastQ format. Both cDNA library construction and Illumina sequencing were performed by NovoGene (Beijing).

Data filtration and de novo assembly

We first filtered the raw reads by removing the adapter sequences, reads with unknown bases call (N) more than 5%, and low quality sequences (<Q20) using an in-house workflow of Novogene and Trimmomatic [34]. Then we removed reads that were likely derived from contaminants of human and Escherichia coli genomes using Bow tie [35]. De novo assembly of clean reads was carried out using a strategy of multiple K-mer lengths and coverage cut-off values. Five different K-mer lengths (21, 31, 41, 51, and 61) and six coverage cut-off values (2, 3, 6, 10, 15, and 20) were used to generate 30 raw assemblies by ABYSS [36]. The raw assemblies were merged to produce a combined assembly. Then CD-HIT-EST [37] was used to eliminate redundancies with sequence identity threshold of 1.0 and word length of 8. CAP3 [38] was used to integrate sequence overlaps with default parameters for three times. A final assembly was generated after removing contigs shorter than 200 bp. All raw sequence reads were mapped back to the final assembly to identify variable sites using Bow tie [35] and SAMtools pipeline [39]. The base call that was consistent with the most mapped reads at a variable site was chosen for the consensus sequences using an in-house Python script.

Gene annotation and GO/KEGG classification

We first constructed a reference dataset for gene annotation, because of the lack of genomic information of any Odorrana species. The dataset combined protein data of seven vertebrate species from the ENSEMBL Database [40], including Anolis carolinensis, Danio rerio, Gallus gallus, Homo sapiens, Mus musculus, Oryzias latipes, and Xenopus tropicalis. Assembled sequences were annotated to the reference dataset based on BLAST similarity using BLASTX [41] with an E-value cut-off of 1E-05.

GO categories and KEGG pathways were used to classify the functions and metabolic pathways of the transcripts. In order to exclude the interference from alternative splicing of transcripts, we first clustered all transcripts that matched the same reference gene; then we removed redundant transcripts and only preserved the longest transcript from each cluster to represent a unique gene. GO and KEGG classification was performed using the Blast2GO [42] pipelines with the default parameters.

AMPs and molecular markers

In order to identify AMPs in the transcriptome of green odorous frog, we blasted the assembled transcripts against the known AMPs from Database of Anuran Defense Peptides (DADP) [43] using Blast-2.2.26+ [41] with the similarity cutoff of 80%. SNP sites were identified using SAMtools [39] pipeline after mapping all clean reads to the assembled transcripts using Bow tie with default parameters [35]. Microsatellite DNA loci were identified by QDD2 pipeline [44], and the same pipeline also automatically designed all associated primers.

Ethics Statement

The animal specimens were collected legally. All animal collection and utility protocols were approved by the Chengdu Institute of Biology Animal Use Ethics Committee.