Background

The internal transcribed spacer 2 (ITS2) rDNA locus has been widely used as a marker for species identification in both free-living and parasitic nematodes for many years [1–3]. Nematodes, as with other invertebrate groups, often exist in large and complex communities. Consequently, deep amplicon sequencing approaches have a potentially powerful role for the investigation of nematode communities similar to the use of bacterial 16S rDNA amplicon sequencing in microbiome studies. For example, the ITS2 rDNA locus has recently been used for “nemabiome” sequencing of parasitic nematode communities inhabiting the gastrointestinal tract of cattle [4, 5]. In that case, reliable species identification was achieved using a small bespoke, curated ITS2 rDNA database of the major relevant cattle gastrointestinal nematode species. However, the wider and more versatile application of deep amplicon sequencing approaches to nematode research will require a more comprehensive, and regularly updated, ITS2 rDNA database equivalent to that available for studying fungi [6]. A eukaryotic ITS2 database has been previously published [7] but has not been updated since 2015 and contains only 1347 sequences in the Nematoda phylum. To our knowledge no other good ITS2 databases exist for nematodes. In this paper, we describe the development of a nematode ITS2 rDNA database and the software to create ITS2 databases for any taxonomy.

Implementation

The nematode ITS2 database was constructed using markerDB, which we have provided as an open-source tool to quickly and reliably construct an ITS2 database for any NCBI taxonomic level. This tool is made available to facilitate reproducibility and transparency, and to provide users with the option to construct their own databases. markerDB is implemented in the R programming language [8] and run as a Snakemake [9] pipeline. The software will run on Linux or MacOS and is dependency free with the use of Bioconda [10]. A brief description of the pipeline follows.

Potential ITS2 sequences are retrieved from NCBI using the rentrez R package [11] based on a text search that will find ITS2 annotated sequences that are limited to the provided taxonomy. The full taxonomy of the downloaded sequences is obtained with the taxize R package [12, 13]. Only taxonomies that are complete with all ranks (Superkingdom, Kingdom, Phylum, Class, Order, Family, Genus, Species) are retained. Additionally, taxonomies with incomplete species names, which contain numbers or ‘sp.’ are removed.

Many of the ITS2 annotated sequences also contain the partial or full upstream and downstream 5.8S and 28S genes and so trimming to the ITS2 region is required. However, no good sequence models exist that capture a wide range of diversity due to the divergence of this region. For this step Infernal, (specifically cmscan) [14] is used to identify the 5.8S and eukaryotic LSU (large subunit, 28S) genes. The co-variance models used to create the nematode database and also provided with markerDB, were taken from Rfam [15]. If a hit to the 5.8S is identified (partial hits allowed) this region and everything upstream is trimmed off. This is repeated downstream for any 28S hits. Any retrieved sequences that do not have hits to either rRNA gene are assumed to be solely ITS2 and are retained in the database. This option can be changed when running the pipeline. A similar strategy has been previously used to identify ITS2 sequences [16] however, that approach relied on custom HMM models to locate the ITS sequences which make it challenging keep the database up-to-date. Using publicly available and frequently updated sequence models from Rfam ensures the most current data is used.

Finally, sequences too long or too short (700 bp and 100 bp, respectively, as set in the configuration) are discarded. The final sequence set contains a fair bit of redundancy and so a non-redundant version of the database with unique sequences only is returned. If an alignment is required an option to align the sequences using MAFFT [17] is also provided but it should be noted that aligning ITS2 sequences from diverse organisms is difficult due to the heterogeneity present. In general we recommend taxonomy assignment methods that do not depend on alignments, particularly for databases covering a large taxonomic range.

The output of markerDB is a fasta file with the final sequences and a corresponding tab-delimited text file with the taxonomy, linked by Genbank accession number. The pipeline also provides function to write out the database in formats used with popular taxonomy assignment methods including dada2 [18], the rRDP Bioconductor package [19], mothur [20], and IDTAXA [21]. For example the IDTAXA output files can be used with our recommended nemabiome analysis workflow (www.nemabiome.ca). A simple shiny app is also provided that allows users to work with the database interactively, filtering taxonomic groups as needed and downloading the filtered data in any of the above formats.

Results and discussion

The database (version 1.0.0 at the time of writing) currently contains 8630 non-redundant sequences with a median length of 263 bp and standard deviation of 97 bp. There are 1429 species and 325 genera and across the taxonomic ranks we were able to obtain good quality, non-redundant sequences for approximately 30% of the taxa in the NCBI database in the Nematoda phylum (see Fig. 1). We have also included an example of data as an illustration of the use of the ITS-2 rDNA database (Additional file 1: Figure 1).

Fig. 1
figure 1

Database size. Number of taxa in version 1 of the ITS2 database compared to the total in the NCBI taxonomy database

Full size image

We have also provided a simple web app, which allows users to search and filter the database and create versions customized to their research area of interest. Further, the database will be updated every 3-6 months, feasible due the automation and reproducibility of the database construction using markerDB. Rapid updates allow researchers to generate analysis that reflect the most current sequences in Genbank.

Conclusions

In conclusion, we provide a database of nematode ITS2 sequences that greatly expands the range of sequences suitable to study both parasitic and free-living nematode communities allowing a broader selection of hosts and environments to be studied. We have also provided open source software to easily and reproduceably build ITS2 databases for any taxonomy of interest.

Availability and requirements

  • Project name: markerDB

  • Project home page: https://github.com/ucvm/markerDB

  • Operating system: Unix-based, i.e. MacOS or Linux

  • Programming languange: R and Python

  • Other requirements: Bioconda

  • License: MIT

  • Any resitriction ot use by non-academics: None