Mature acorns from Holm oak (Q. ilex L. subsp. ballota [Desf.] Samp.) were collected from a tree located in Aldea de Cuenca (province of Córdoba, Andalusia, Spain). Acorns were germinated and seedlings grew in a chamber under controlled conditions (a 12 h photoperiod, a temperature of 21 ± 1°C, a relative humidity of 60 ± 5% and an irradiance of 200 μmol m⁻² s⁻¹, Echevarría-Zomeño et al., 2009). Germinated embryo, leaves and roots from 1 year plantlets were collected separately, weighted, and individually frozen in liquid nitrogen. The plant material used for RNA sequencing experiments consisted in a pool generated by mixing equal amounts of homogenized tissue from acorn embryo, leaves, and roots.

Total RNA was extracted from 50 mg pooled plant sample according the procedures previously set up in our laboratory for Q. ilex samples (Echevarría-Zomeño et al., 2012). Contaminating genomic DNA was removed by DNase I (Ambion) treatment. Total RNA was quantified spectrophotometrically (DU 228800 Spectrophotometer, Beckman Coulter, TrayCell Hellma GmbH & Co. KG). The high quality and integrity of the RNA preparation was tested electrophoretically (Agilent 2100 Bioanalyzer). Only high-quality RNAs with RIN values > 8 and A₂₆₀:A₂₈₀ ratios near 2.0 were used for subsequent experiments.

The library construction of cDNA molecules was carried out using Illumina TruSeq Stranded mRNA Library Preparation Kit according to manufacturer instructions using 2 μg of total RNA followed by poly-A mRNA enrichment using streptavidin coated magnetic beads and thermal mRNA fragmentation. The cDNA was synthesized, followed by a chemical fragmentation (DNA library) and sequenced in the Illumina Hiseq 2500 platform, using 100 bp paired-end sequencing (Conesa and Götz, 2008; De Wit et al., 2012).

The raw reads obtained from the sequencing platform were pre-processed in order to retain only high-quality sequences to be subsequently used in the assembly. Thus, each original sequence was quality trimmed considering several parameters (quality trimming based on minimum quality scores, ambiguity trimming to trim off e.g., stretches of Ns, base trim to remove specified number of bases at either 3′ or 5′ end of the reads). The pre-processing parameters used were selected as following: trimming sequences by maximum 2 ambiguous nucleotides), minimum mean quality assuming error probability < 0.01, and filtering out those sequences shorter than 30 nucleotides.

Three different assemblers were employed to de novo assemble the Q. ilex transcriptome, considering there is not a reference genome available, and further evaluated to contrast the results obtained (Figure 1).

Trinity 2.4.0. performs a de novo assembly using an algorithm based on Bruijn graphs (Grabherr et al., 2011). For the assembly, Trinity 2.4.0 was launched with a k-mer value (k = 25).

Ray 2.3.1. assembly uses de Bruijn graphs but its framework is not based on the Eulerian steps. Specific subsequences, seeds, are defined, and for each of them, the algorithm extends it to a contig. Heuristics are defined that control the extension process in such a way that the process stops if, at some point, the readings family does not clearly indicate the address of the extension (Boisvert et al., 2010). In this case we selected a k-mer value of 31.

MIRA 4.9.6 software (Chevreux et al., 1999), unlike Trinity and Ray, is based on the strategy known as Overlap /Layout/ Consensus. Following the author guidelines/recommendations for Illumina data, we used the complete raw data without a filtering process like we described previously.

Evaluation of the structure of the generated assemblies was done with the QUAST software (Gurevich et al., 2013).

The assemblies obtained using the three aforementioned softwares were blasted (e-value of 10⁻³⁰) against the most accurate and nearest phylogenetic transcriptome currently available, the oak transcriptome (containing Quercus robur and Quercus petraea sequences) (Lesur et al., 2015). That transcriptome database is divided in two files OCV3_91K and OCV3_101K but OCV3_91K has a larger amount of valuable information of Quercus spp. transcriptome. So, we chose OCV3_91K as a general oak transcriptome database.

There are differences between the three assembled transcriptomes in terms of transcriptome architecture/structure. Thus, the N50 value, number of contigs and the average length of the sequences generated by each algorithm differ (Table 1).

Considering these results, we can state that MIRA generated more and longer contigs than RAY and Trinity (MIRA>RAY>Trinity), suggesting that a more robust architecture/structure is obtained by MIRA for the Q. ilex transcriptome assemby. Upon the continuous development of NGS methods, data processing, and transcript assemby remains a main challenge. Several studies have been published devoted to evaluate different de novo assemblers varying in performance and quality in terms of number and length of transcripts and computational speed (Clarke et al., 2013). Besides, it has been reported that the quality of the assembly using a given software depends on the biological sample on study (Bradnam et al., 2013). Thus, these aspects should be taken into consideration when comparing different softwares.

The comparison between the sequences generated from Q. ilex and those available from the close species, oak transcriptome, reveals that MIRA assembly was the one which shared the higher number of transcripts (73073), followed by RAY assembler (Table 1). Besides, MIRA assembly sequences blasted against oak transcriptome render the longest alignment lengths and better blast scores (Figure 1).

Taking into consideration the data and parameters evaluated (Table 1 and Figure 1), we decided to use the MIRA assembly to continue with the corresponding annotation of Q. ilex transcriptome. After blastX was completed against Uni-Prot (Swiss-Prot) curated database (e-value of 10⁻⁵), followed by the corresponding mapping process, 31973 annotated sequences were obtained by Blast2GO (Conesa and Götz, 2008).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.