De Novo transcriptome assembly and annotation of the wild service tree Sorbus torminalis (L.) Crantz

Abstract

Objectives

The wild service tree, Sorbus torminalis (L.) Crantz, is a sub-Mediterranean forest tree species native to Europe, Asia Minor, Caucasus and north-western Africa. Despite its ecological and societal importance, publicly available molecular resources for this species are scarce, limited to the chloroplast genome. Within the European Project FORGENIUS (Improving access to FORest GENetic resources Information and services for end-Users, Grant Agreement ID: 862221) we assembled the transcriptome of the wild service tree with the aim to use it as reference to develop new genetic markers and perform a large-scale population genetics study in this species.

Data description

An RNA-seq library was generated from pooled RNA extracted from four different tissues (young leaves, mature leaves, stems and roots) of a single S. torminalis sapling and sequenced on the Illumina HiSeq 2500 platform. This dataset was used to produce and annotate the first transcriptome assembly for S. torminalis, paving the way for future genomic research and applications in this species. For example, this study will facilitate the development of molecular markers such as SSRs and SNPs, useful for genetic studies, breeding programs, and conservation efforts and will serve as a reference for future differential expression studies, providing a deeper understanding of the species’ biology.

Objective

Rare and scattered forest tree species are important components of forest biodiversity and represent a key option for diversifying simplified forest ecosystems and eventually increasing their resilience. It is often difficult to monitor the adaptive potential of these species because, contrary to more widespread, ecologically and economically relevant forest tree species, the resources available for their genomic characterization are scarce. This is the case for the wild service tree Sorbus torminalis (L.) Crantz (Rosaceae), a diploid (2n = 34), insect-pollinated and animal-dispersed species native to Europe, Asia Minor, Caucasus and north-western Africa. S. torminalis has historically been used in various applications in the fields of medicine and nutrition and produces durable wood of high economic value [1, 2]. To date, most of the genetic studies on this species are based on a limited number of microsatellite markers [3,4,5] and, more recently, the complete chloroplast genome sequence of the species was released [6]. In the framework of the European Project FORGENIUS (Improving access to FORest GENetic resources Information and services for end-Users, Grant Agreement ID: 862221), we assembled and annotated for the first time the transcriptome of the wild service tree. Transcriptome assembly will serve as a reference for developing a new genotyping tool for S. torminalis using Single Primer Enrichment Technology (SPET), a recently developed targeted genotyping-by-sequencing technology [7]. This genotyping tool will enable to explore the distribution of genetic diversity at a genome-wide scale in this species, allowing forest geneticists to further investigate the adaptive responses of the species to environmental changes.

Data description

The RNA sequencing data used to construct the transcriptome assembly of Sorbus torminalis were generated from total RNA extracted from four different tissues (young leaves, mature leaves, stems and roots) of a single S. torminalis sapling provided by the public nursery of Veneto Agricoltura (Montecchio Precalcino, VI, Italy). The RNA from the different tissues was pooled in equimolar amounts and used to prepare the RNA-Seq libraries that have been sequenced on the Illumina HiSeq 2500 platform as 250 bp paired-end reads. The raw Illumina RNA-seq data were deposited in the NCBI Sequence Read Archive (SRA) database with ID SRP489964 (Table 1, Data file 1). The methodological details of the RNA sequencing and bioinformatics analysis are reported in Data File 2 (Table 1). The quality of the RNA-seq data was evaluated using FastQC v0.11.3 before and after the trimming step (Cutadapt v1.16 [8] and ERNE_FILTER v1.4.6 [9]). The quality control metrics of the trimmed reads were aggregated using MultiQC v1.6 [10]. Data File 3 (Table 1) shows the distribution of the mean quality scores at each read position and the mean read counts per quality score for both first-in-pair and second-in-pair reads. All positions in first-in-pair reads and almost all positions in second-in-pair reads were characterized by an average Phred score of at least 33. A total of 244,607,284 clean read pairs (approximately 91% of the raw reads) were retained after quality trimming and filtering.

The filtered reads were assembled into 71,309 unigenes with an average length of 543.1 bp and N50 of 658 bp using the de novo assembler Trinity v2.15.1 [11]. The transcriptome assembly has been deposited in the Transcriptome Shotgun Assembly database of NCBI under accession number GKRU00000000 (Data file 4, Table 1). Using TransDecoder v5.0.1 (https://github.com/TransDecoder/TransDecoder), a total of 28,662 open reading frames (ORFs) were predicted, of which 6,737 had both start and stop codons, while the remaining ORFs contained only partial coding regions (Data files 5 and 6, Table 1). Functional annotation of the unigenes was performed using BLAST2GO version V.6.0.3 [12] on BLASTx [13] results against the Viridiplantae section of the GenBank non-redundant protein database. Approximately 73% of the unigenes had a BLASTx hit to proteins included in the database. Most of them matched proteins belonging to the Malus or Pyrus genera (Data file 7, Table 1). Using Blast2GO, we assigned a total of 162,340 gene ontology (GO) terms to 43,240 of the 52,129 unigenes for which we obtained a BLASTx match. Most of the assignments (69,523; 42.8%) belonged to the ‘Molecular Function’ category, followed by those belonging to the ‘Biological Process’ (52,926; 32.6%) and ‘Cellular Component’ (39,891; 24.6%) categories. A total of 78,905 Enzyme Codes (EC) were assigned to 21,518 different unigenes. The distribution of the most frequent gene ontology (GO) terms in the three different GO categories and of the most represented EC is shown in Data files 8 and 9 (Table 1) respectively, while the complete annotation is available in Data file 10 (Table 1).

Table 1 Overview of data files/data sets

Limitations

The transcriptome assembly was obtained from RNA extracted from four tissues at a single time point in a single individual, and therefore it does not comprise an exhaustive set of all transcripts of this species but only those constitutively expressed under such specific conditions.