GreeNC 2.0: a comprehensive database of plant long non-coding RNAs

Abstract

The Green Non-Coding Database (GreeNC) is one of the reference databases for the study of plant long non-coding RNAs (lncRNAs). Here we present our most recent update where 16 species have been updated, while 78 species have been added, resulting in the annotation of more than 495 000 lncRNAs. Moreover, sequence clustering was applied providing information about sequence conservation and gene families. The current version of the database is available at: http://greenc.sequentiabiotech.com/wiki2/Main_Page.

INTRODUCTION

Long non-coding RNAs (lncRNAs) used to be considered as transcriptional noise in the past decades, but lately, this class of molecules has gained increasing attention in epigenetic research and it is now recognized to have an important role in mediating the transmission and the expression of genetic information (1). LncRNAs are RNA molecules longer than 200 nucleotides with no protein-coding ability (2), despite this, they are involved in fundamental biological processes, and their activities are complex and diverse. In fact, lncRNAs could help in the regulation of protein modification but also chromatin remodelling, RNA metabolism, transcription, DNA methylation, and many other processes (1). Due to their activity in a wide number of pathways, lncRNA are very well studied in human and clinical applications (3). In plants, many lncRNAs have been characterized in model organisms as Arabidopsis thaliana, Zea mays, and Solanum lycopersicum. For instance, Arabidopsis lncRNA APOLO (AUXIN-REGULATED PROMOTER LOOP) regulates the expression of un-related distant auxin-responsive genes during the lateral root development by modulating local chromatin conformation (4). Zea mays PILNCR1 is involved in the plant adaptation to phosphate deficiency (5). LncRNA1459, detected in Solanum lycopersicum, has been shown to be involved in the fruit ripening process (6).

In order to help the scientific community to study plant lncRNA sequences and functions, we developed and published the GreeNC database in 2015 (7). Since then, the database has been accessed >250 000 times and it has become a reference for the plant scientific community working on lncRNAs. In the last years tens of new plant species have been sequenced and for many of the species shown in GreeNC new or updated reference genomes have been published, for this reason we present GreeNC 2.0 a new update where lncRNAs from new 78 species were added and 16 species were updated. In addition, we performed an extensive sequence clustering in order to detect orthologous groups of lncRNAs both between species and within species. With this additional information researchers will be able to detect whether candidate lncRNAs belong to gene families and if they are conserved across species.

MATERIALS AND METHODS

Genome and annotations

FASTA sequences of transcripts were downloaded from Phytozome v13 and Plants Ensembl version 51. The assembly version of each species is reported in Supplementary Table S1. Only un-restricted genomic data were used (8–85). For Oryza and for Brassica rapa, transcripts were downloaded from Plant Ensembl 51 and Phytozome v13, respectively.

Identification of lncRNAs

As in the previous version of GreeNC (7), two bash scripts were used to identify lncRNAs among the downloaded transcript sequences (Figure 1). With the first script coding potential is calculated maintaining only transcripts with a minimum length of 200 nucleotides and an ORF shorter than 120 amino acids by using Ugene (38.1). Sequences were blasted (v2.9.0) against SwissProt (2021/04). CPC (0.9-r2) was used to assess the protein-coding potential of transcripts. To discriminate other non-coding transcripts from lncRNAs, and to identify possible miRNA precursors, a second script was used. Transcripts were analyzed by cmscan (Infernal 1.1rc4) against the RFAM database (release 14.6). BLASTn (2.6.0) was used against a database of mature plant miRNA sequences from miRBase (release 22.1). The final list of lncRNAs was divided into high- and low confidence. Transcripts without hits in BLASTX described as non-coding by CPC and not considered as miRNA precursors, were classified as high-confidence lncRNAs. Those without hits in BLASTX but considered coding by CPC, those with BLASTX hits considered noncoding by CPC, and those considered miRNA precursors, were marked as low-confidence lncRNAs. To exclude putative transposons, RepeatMasker was used, in order to identify transcripts containing predicted repetitive regions. These transcripts are also classified as low-confidence. RepeatMasker (4.1.0) was executed with a custom library obtained by RepBase (86) with the following parameters -no_is, -gff, -nolow.

Figure 1.

Overview of the in-house developed computational pipeline for lncRNA annotation, which consists of script 1 (A) and script 2 (B).

Relational database

Data was imported into a MySQL-based relational database stored on an Ubuntu server (Ubuntu 18.04.4 LTS). This database was then integrated into a MediaWiki by mapping relational data fields against predefined templates via Semantic MediaWiki. Transcript sequences in a FASTA file were formatted using makeblastdb. Sequence retrieval is based on blastdbcmd. An Express Node.js API web service was created to expose both sequence retrieval and BLAST searches via client JavaScript from the MediaWiki interface.

OrthoFinder

To evaluate sequence similarity and cluster lncRNAs in orthogrups, an OrthoFinder (87) analysis was executed with the following parameters -d, -f, -S diamond_ultra_sens. As input files, lncRNA sequences from all the species were used.

RESULTS

The previous version of GreeNC (7) included 43 species, resulting in a total of 120 000 annotated lncRNAs. After this update, GreeNC 2.0 includes information on >495 000 transcripts from 94 species between plants and algae (Figure 2). More than 327 000 transcripts were annotated as high confidence lncRNA. With this update, the highest percentages of lncRNAs were annotated in Triticum dicoccoides (7.7%), and Aegilops tauschii (6.9%) and Hordeum vulgare (4.8%), while the lowest in Juglans regia (0.13%), Chara braunii (0.12%) and Cyanidioschyzon merolae (0.02%).

Figure 2.

A snapshot of a Cucumis melo entry from the GreeNC database. (A) Header, to navigate through the website and access to the tools and the pages of the species; (B) table of gene information reporting genomic coordinates, genome version, the source of the genome assembly and if the gene encodes at least one coding transcript; (C) table of transcript features reporting the kind of lncRNA (low-/high-confidence), if it is a precursor of miRNAs, length, orthologous group, sequence and links to get the Open Reading Frame (ORF), the Coding Potential, the folding energy and the GC content; (D) an optional table that provides links to other databases, when applicable, and giving information about the version of the database and the e-value of the match; (E) table of transcripts belonging to the same orthogroup reporting the kind of lncRNA, length, folding energies (AMFE, MFEI), GC content (F) a schematic representation of the gene and transcript models.

Even if it is known that lncRNAs do not show high conservation at nucleotide level (88), we decided to perform a sequence clustering based on the Orthofinder algorithm in order to provide information about highly conserved lncRNAs. About 39% of the 542 656 identified transcripts were assigned to orthogroups. In total, 65 191 orthogroups were identified however, as expected, no orthogroups were present in all the species. Despite this, shared orthogroups were identified between species of the same genus, suggesting the presence of genus-specific lineages of lncRNAs (i.e. Triticum, Arabidopsis thaliana, Oryza, Gossypium, Brassica). Moreover, the presence of species-specific orthogroups highlights that long non-coding transcripts may be organized in gene families.

A total of 24 743 orthogroups were identified as species-specific, with a mean of 242 orthogroups per species. The highest number of species-specific orthogroups was recorded in Triticum dicoccoides (3487 orthogroups), while the lowest was detected in Cyanidioschyzon merolae (2). A total of 81 446 transcripts (15% of the total) were classified as species-specific, with a mean of 798 transcripts per species. Also in this case, the highest and lowest values were recorded in Triticum dicoccoides (17 234 transcripts) and Cyanidioschyzon merolae (4), respectively.

DATA AVAILABILITY

The GreeNC database is a MySQL relational database and it is freely accessible at: http://greenc.sequentiabiotech.com/wiki2/Main_Page.

The pipeline for lncRNA prediction is available at: https://github.com/sequentiabiotech/GreeNC.

SUPPLEMENTARY DATA

Supplementary Data are available at NAR Online.

FUNDING

European Union's Horizon 2020 Research and Innovation Program under the Marie Słodowska-Curie [101007438]. Funding for open access charge: Self funding.

Conflict of interest statement. None declared.