Mining viruses in public databases unveils the diversity within the Deltaflexiviridae family

doi:10.21203/rs.3.rs-7753018/v1

Download PDF

Research Article

Mining viruses in public databases unveils the diversity within the Deltaflexiviridae family

https://doi.org/10.21203/rs.3.rs-7753018/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Cloud computing platforms aided the scalability and applicability of viral mining in genomic databases. The Serratus project reported SRA accessions that may contain viral sequences. This study analyzed SRA accessions that contain sequences with similarity to members of Tymovirales to verify the existence of viral genomes. All steps in the genome mining analysis were conducted by a pipeline running in virtual machines hosted on the Google Cloud platform. Manual curation of the pipeline output obtained 111 putative genomes. Among the found genomes, four were classified as isolates within the Betaflexiviridae family, and two were putative new members of the Alphaflexiviridae family. Another four sequences were classified into the Emraviridae family, which is still not accepted by the ICTV. The phylogenetic reconstruction of the Deltaflexiviridae family indicates the formation of three distinct clades, one of them containing 81 genomes (34 putative new species), a second clade with 18 novel genomes (7 putative new species), and a third with one putative new species. Given the high divergence found between these three groups, we suggest the establishment of a new family, Epsilonflexiviridae, and the split of the Deltaflexiviridae family into two by establishing the family Thetaflexiviridae. The results of this work give insight into important aspects of the evolutionary history of the order Tymovirales and offer new ways for virus-mining projects in genomic databases.

The rapid growth of the Short Read Archive (SRA), a database that stores raw data from genomic sequencing experiments [1], sparked interest in using this information to survey new viral species [2, 3]. Such efforts contributed to the expansion of the diversity borders of plant viruses [4–7] and viroids [8] with significant economic interest and helped unveil the diversity in groups of previously neglected viruses [7, 9–11]. However, the sheer volume of information presents data processing challenges. Nevertheless, these hurdles can be overcomed by leveraging cloud computing, which offers commercially available, massively distributed computational resources.[12]. The mining process requires multiple intense computational tasks to be repeated for all the accessions under study, which prompts the need for data processing automation. These steps are often sequential, with the output of one step fed into the next program, which facilitates the organization of such steps into automated computational pipelines [13]. The construction of genomic pipelines is an area of intense research, and several pipelines are available for virus discovery [14–18]. Although extremely useful and easy to use, it is often very challenging to modify the available automated pipelines for the use of applications outside their intended scope. A possible solution is to use pipeline platforms that offer flexibility and reproducibility [19, 20], combined with straightforward integration with cloud computing resources [19]. These platforms often have a very active community of collaborators who share their reproducible pipelines for general use, combining flexibility with a less steep learning curve [21]. Still, the sheer volume of data is an obstacle in itself. However, this barrier has also been reduced by the Serratus project, a resource recently proposed [22]. Through an efficient cloud computing infrastructure on Amazon Web Services (AWS), the authors successfully pre-aligned numerous accessions, identifying those likely containing virus sequences [22].

The family of viruses Deltaflexiviridae was proposed as a member of the Tymovirales order, with a genome length of approximately 8kb, featuring a poly(A) tail at the 3’ end [23]. The family was proposed based on a virus isolated from the fungi Sclerotinia sclerotiorum, a plant pathogen [25], which has been the target for virome studies [26, 27] due to its high economic impact on crops [25]. This new family was included in the list of the currently accepted families within the Tymovirales order: Alphaflexividae, Betaflexiviridae, Gammaflexiviridae, and Tymoviridae [24]. The genome organization varies considerably, with the only gene present in all the members coding for a polyprotein. Some viruses contain other ORFs that do not encode the polyprotein; the functions have not been elucidated yet, so they are all annotated as hypothetical proteins. These ORFs are named using a consistent convention, with numbers (HP2, HP3, HP4, and HP5) appended to the "HP" prefix, which stands for hypothetical protein.

With the aid of the Serratus resource, we investigated candidate accessions that contained possible viral genomes belonging to the family Deltaflexiviridae. Our findings reveal tens of possible new viruses and shed light on the evolution of the order Tymovirales as a whole.

Assembly pipeline

The mined viral sequences described in this work were obtained using an automated pipeline running on Google Cloud servers that takes a list of SRA accessions (Supplementary Table 1) as input and returns probable viral contigs and coverage statistics (Fig. 1). The accessions with likely viral sequences were downloaded from the Serratus interface and uploaded to a virtual machine on Google Cloud with 200 GB of storage, 32 threads, and 64 GB of RAM.

The first pipeline step is to download the raw reads from the SRA database, which is available from Google Cloud servers. Thus, when available, the raw reads were simply copied from the cloud, eliminating the need for data transfer. When this was not possible, the reads were acquired through the ENA toolkit due to its faster download. Accessions unavailable through the ENA database [28] were downloaded by the SRA toolkit. Metadata associated with the SRA accession was acquired through EUtilities [29]. SPAdes was used to assemble the downloaded reads, with the rnaviralspades option [30] enabled and all other options kept at their default settings. In cases where SPAdes returned errors, Megahit replaced it (Fig. 1), also with default settings [31]. The accessions referring to reads that could not be assembled by either assembler were discarded. Contigs with a length below 500bp were discarded, and the remaining were subjected to the filtering step. Given that the focus is on the Deltaflexiviridae family, the assembled contigs were filtered by a BLAST search [32] against a custom database containing amino acid sequences of viruses belonging to this family deposited in public protein databases. After the filtering step, the raw reads were mapped to the filtered sequences using BBmap [33] to calculate coverage statistics.

Manual curation and annotation

The virus and virus-like genomic sequences were manually curated in the Geneious software [34]. ORFinder was used to predict putative ORFs with only ATG as a start codon and a minimum ORF length of 150bp. The curation process took into account genome composition; that is, predicted open reading frames were compared with those found in the available genomes. An iterative extension protocol was then applied to genomes with large missing portions. The incomplete genomes were used as references for mapping against the unfiltered contigs using Geneious [34]. To extend the genome, the non-overlapping portion of the alignment was added to the reference only if the overlapping mapped contigs contained no more than two mutations. Viral sequences successfully extended using this method were remapped against the raw reads with BBmap [33], and coverage statistics were calculated. ORFs predicted with ORFinder built into Geneious were manually curated by removing internal and reverse coding frames. To annotate the ORFs, information from an orthologous search performed by OrthoMCL [35] and conserved domains predicted by InterProScan [36] was combined to curate the ORFs. Additional annotation of the clusters was assigned through manual searches using the web version of the InterProScan [37]. The annotated polyprotein ORFs were then extracted for further analysis. Pairwise identities between the polyprotein amino acid sequences were calculated by SDT [38] using Muscle [39] as the aligner. The phylogenetic reconstructions of the polyprotein sequences were estimated by IQ-Tree [40], which includes a step to choose the best-fit substitution model. The sequences were aligned by MAFFT [41] and trimmed using the “automated1” option in TrimAL [42]. The online tool iTOL [43] was used to draw the phylogenetic trees and genome organization. The summary figures were plotted using the Python package plotly.

Possible new members in the Tymovirales order

Initial similarity searches indicated that the new sequences probably belonged to members of the Tymovirales order. To elucidate the placement within this order, a dataset containing members of the order is used to construct the phylogeny of the group. This dataset combines reference genomes from the Alphaflexiviridae, Betaflexiviridae, Deltaflexiviridae, and Tymoviridae families with sequences obtained through an NCBI database search [44]. Additionally, a broader BLAST similarity search was performed against a database containing all the proteins deposited in the Virus database in the NCBI database [45]. After filtering hits that generated shorter alignments, the corresponding genomes were also added to the dataset. The polyprotein amino acid sequences were then extracted, and the ones containing very short or truncated ORFs were excluded from the analysis. The resulting phylogenetic tree is shown in Fig. 2. The highlighted clades indicate where the mined viral sequences clustered.

The phylogeny of the members of the Tymovirales order corroborates the family classification currently accepted for the well-known plant-infecting viruses such as the Alphaflexiviridae, Betaflexiviridae, and Tymoviridae, as they form monophyletic clades (shown in green, orange, and grey in Fig. 2, respectively) with bootstrap values above 90. Additionally, the phylogenetic tree corroborates the subfamily structure within the Betaflexiviridae family, with the two subfamilies Quirivinae and Trivinae forming well-supported clades (shown in orange in Fig. 2). Another well-defined clade, shown in blue in Fig. 2, with high statistical support (Bootstrap above 90), included members that have been suggested to constitute a new family called Emraviridae [46]. Surprisingly, the clade that would form the Deltaflexiviridae family presented an unexpected structure: Two large clades (shown in red and purple in Fig. 2), harboring most of the new sequences (100 sequences out of the 111), displayed considerable divergence from each other, despite stemming from a common ancestor and forming a monophyletic clade. The evident separation between these two clades, as indicated by the length of the branch marked in purple, suggests that the Deltaflexiviridae can be divided into two families, proposed in this work as Deltaflexiviridae and Thetaflexiviridae. Two other sequences described in this work formed a separate clade (shown in pink in Fig. 2). Given their high divergence from the rest of the clade B, these two sequences should also be classified as a new family, proposed in this work as Epsilonflexiviridae.

Two novel sequences clustered with Alphaflexiviridae (marked in green in Fig. 2), and one of them, hereby named East River virus 2, diverged from the common ancestor for the whole family. A second phylogenetic tree, containing a subset of the sequences in Fig. 3, created by sampling three members of each Alphaflexiviridae family and using the Tymoviridae member turnip yellow mosaic virus as the outgroup, shows that East River virus 2 probably belongs to this family, despite sharing only the polyprotein ORF with the other members of the family (Fig. 3B). The phylogenetic tree also places the other novel virus, East River virus 1, among the members of the Allexivirus genus (Fig. 3A). This grouping is well supported by the high Fast Bootstrap value and by genomic composition (Fig. 3B), since East River virus 1 shares genes with the other members in the genus. Interestingly, this virus also has a coat protein gene ORF overlapping with the 40kDa protein gene ORF, and it is also longer than its counterparts.

The Betaflexiviridae family contained four new sequences, which were highly similar to published sequences (Fig. 4). Repeating the approach used for Alphaflexiviridae, three reference sequences were selected from each genus to place the new putative viruses. The high similarity indicates that these viruses are isolates from the existing species belonging to the genus Carlavirus. Three isolates originated from samples taken from Gansu. They belonged to the potato virus S (closest BLAST hit identity: AAP76207–98.22%), potato virus H (closest BLAST hit identity: AEI55831–96.1%), and potato virus M (closest BLAST hit identity: YP_277428 − 94.68%). A fourth isolate identified as a member of the potato virus M (closest BLAST hit identity: UTQ50775–98.32%) originated from a sample of wintersweet, Chimonanthus praecox (Fig. 4).

Four of the new genomes clustered with the isolated group, suggesting they can be classified into the new family Emraviridae (shown in blue in Fig. 2). A phylogenetic tree (Fig. 5A) with fewer sequences and, thus, a higher resolution indicates that these four genomes (shown in bold and italic in Fig. 5) belong to new species within this family. The highest identity to a BLAST hit among the new sequences is 43.4% (AQM32763), thus corroborating the classification of the novel putative genomes as new species.

As this family was recently proposed, it is yet to be recognized by the ICTV, consequently, there is still no genus-level classification. Yet, there are some indications of a possible genus classification. A combination of orthologous search clustering and phylogenetic analysis revealed the formation of five well-defined groups, corroborated by both genome organization clustering and tree structure. Three out of the four new viruses clustered with clade V, which is also characterized by the presence of HP1 (shown in green in Fig. 5B). The remaining genome clustered within clade I, which is corroborated by the presence of HP3, even though not all members of the clade contain this gene.

Deltaflexiviridae family split

Interestingly, most of the sequences described in this work were grouped with known members classified within the family Deltaflexiviridae. However, the group exhibits a remarkable topology, characterized by three distinct and clearly defined clades (Fig. 2). Additionally, the published sequences classified as members of the family did not form a monophyletic clade. For example, Sclerotinia sclerotiorum deltaflexivirus 1 (Fig. 6), the type species for the Deltaflexiviridae family, and Fusarium deltaflexivirus 2 (Fig. 7) are categorized within the same family, yet they cluster in distinct clades.

The high divergence between the clades suggests that these three clades should be classified into different families. Thus, we propose the creation of two families to represent the evolutionary history of this group more accurately, as estimated by the phylogenetic analysis, namely Epsilonflexiviridae (shown in pink in Fig. 6) and Thetaflexiviridae (Figs. 7 and 8). Despite this phylogeny showing some incongruences with the tree shown in Fig. 2, such as the placement of the Epsilonflexiviridae, the divergence is still consistent, which supports our proposed classification.

Following our proposed classification, the Deltaflexiviridae family will then encompass 18 out of the 111 new sequences. The new Deltaflexiviridae phylogenetic tree is presented in Fig. 6, with the new genomes shown in bold and italics. Interestingly, the genome content is quite diverse within this group. The polyprotein is the sole shared element among all viruses, and seven clusters of orthologs are identified within the group.

Pairwise amino acid identity comparisons revealed that 13 out of the 18 genomes classified into the Deltaflexiviridae family belong to new species (marked with circles in Fig. 6, Supplementary Table 2). However, using the more conservative prerequisite and assuming that only genomes mined from samples originated from a single individual indicated that 7 of the 18 putative viruses should be classified as new species (marked with black circles in Fig. 6). Four novel species, three mined from single plant samples collected from conifer samples (Picea glauca deltaflexivirus, Cupressus duclouxiana deltaflexivirus, Cupressus gigantea deltaflexivirus, and sharing 55.16%, 55.53% and 55.45% identity with QYF50206, respectively) [47, 48], and one from a single lesion caused by the plant pathogen Puccinia striiformis in a wheat leaf [49], clustered together (Puccinia striiformis deltaflexivirus − 55.43% identity with QYF50206). Another cluster containing four new genomes harbored a putative virus, Picea deltaflexivirus, isolated from a single spruce individual (58.01% identity to UYL94495), which clustered with two new genomes mined from a sample of the moss Dicranum scoparium, which could be classified as isolates of a new species, but did not come from a sample of a single individual [50]. Finally, a genome mined from a single-plant sample of the conifer Abies sachalinensis, (Abies sachalinensis deltaflexivirus − 57.7% identity to 3QYF50206), clustered with a genome mined from a soil sample from Antarctica.

In addition to the putative novel species, viruses that belonged to already described species were also detected by our methods. Among those, three genomes were classified as members of the species Eryphisie necator associated deltaflexivirus 4 and Sclerotinia sclerotiorum deltaflexivirus 1, viruses associated with plant-infecting fungi. The isolate originating from corn lacks a part of the genome, indicating that it is incomplete. The remaining new genome within this cluster was classified as an isolate of the already described Leptosphaeria biglobosa deltaflexivirus 2 (Fig. 6).

The family Epsilonflexiviridae (depicted in pink in Figs. 2 and 6) comprises two members. The amino acid identity between Chimonanthus praecox epsilonflexivirus and the closest sequenced entry in GenBank (QDW81317) is 34.33%. The sample from which this putative virus was derived from a single wintersweet plant (47), supporting its classification as a new species. In contrast, information regarding the sample of the other family member, Macrocybe gigantea epsilonflexivirus, isolated from an edible mushroom, could not be located. Therefore, although it shares 33.83% identity with the closest deposited sequence (QDW81317), the classification of this putative virus as a novel species should be approached with caution.

Meanwhile, the family Thetaflexiviridae (shown in purple in Fig. 2) harbored 81 of the 111 new sequences. The novel likely viruses are presented in Figs. 7 and 8 with a higher-resolution phylogeny, marked in bold and italics. To facilitate the visualization, portions of the tree were collapsed; specifically, the highlighted section in Fig. 7 is collapsed in Fig. 8 and vice versa.

Taking the threshold of 90% amino acid identity and the genomic content, 18 novel genomes belong to species already described, and 67 novel genomes probably belong to new species. Taking into account only genomes mined from samples in which a single individual was sequenced, this number drops to 34. Since some of the new genomes assigned as probable new species were closely related (above 90% identity), only 55 new species were predicted to exist under the identity criterion within the Thtaflexividae family. For instance, a putative virus named Crawfurdia deltaflexivirus (55.62% identity with QKN22686), mined from a sequenced museum accession [51], showed high identity with another virus mined from leaves of a single individual of the Chinese cypress [48], Cupressus duclouxiana. Finally, 25 new species fit the single-individual sample criteria within the Thetaflexiviridae family.

A notable large clade contains genomes with a single ORF, the polyprotein (see highlighted section in Fig. 8). Interestingly, a putative new virus, Brassica napus thetaflexivirus, is placed within this group, but contains HP3, RNAse III, HP7, and two more genes with no detected orthologs. It is unclear whether these distinct additional ORFs were acquired independently or if they were present in the ancestor, as this virus was not positioned as diverging from the ancestor of the entire single ORF group.

This clade also harbors 41 of the 81 new sequences characterized in this work. All of them could be classified as new species under the similarity criteria. However, taking into account only viruses mined from the single individual samples, only 30 should be considered to belong to a new species. There is a clade composed only of new species, comprising, for instance, the putative virus Tetracentron sinense thetaflexivirus 2, mined from a sequencing project that sampled leaves from a single individual, [52] and Picea wilsonii thetaflexivirus 2, mined from sequencing data of a single individual of the conifer Pìcea wilsonii [50] (Fig. 8). This highlights the limited understanding of the diversity within this group.

The orthology detection performed by OrthoMCL [35] resulted in five new groups of genes in different viral genomes with likely shared evolutionary history, hereby named hypothetical protein 6 through 10 (HP6, HP7, HP8, HP9, and HP10). This method was also capable of reproducing the orthology relationships previously indicated; that is, orthologs of the hypothetical proteins HP2 to HP5 were also found. However, information about the function of these genes could only be found for the HP2 via InterproScan searches [37], which includes an RNAse 3 domain commonly found in viruses infecting fungi.

Additionally, patterns of gene presence/absence were also found. For instance, the orthology clustering showed that genomes lacking HP7 (yellow in Fig. 6) contained the gene that showed similarity with the protein annotated as hypothetical protein 3 (HP3) (orange in Fig. 6), which could indicate that this cluster of genes and HP3 are orthologous.

The genomic content is more conserved than that observed in the family Deltaflexiviridae. Interestingly, genes annotated as the coat protein gene and HP3 are mutually exclusive. The coat protein gene is shared between the Deltaflexiviridae and Thetaflexiviridae, with the latter found in two genomes sourced from Angelo Coast and Alaska (Fig. 7). Given that the presence of this gene is sparse even within the Deltaflexiviridae, further studies may be necessary to understand its evolutionary history.

A set of 28 detected clusters contained only two genes in closely related viruses, hereby named clade-restricted clusters (dark grey squares in Figs. 6, 7, and 8). Notably, the HP5 gene was also present in only two genomes, the previously described Agave tequilana deltaflexivirus and a newly described putative virus, Spruce root thetaflexivirus. However, HP5 was not assigned as clade-restricted, being another example of a gene shared between the families Deltaflexiviridae and Thetaflexiviridae.

In summary, the five previously described hypothetical proteins (HP2 to HP5) were present in both families. Among the orthologs reported in this work, HP6 and HP9 were common to both families, HP8 was exclusive to Deltaflexiviridae, and HP7 and HP9 were only detected in Thetaflexiviridae.

Pipeline efficiency

The pipeline was executed for 202 accessions available in the Serratus platform. Despite the indications that all of them contained viral sequences, our pipeline failed to detect new genomes in more than half of the accessions. The reads from two accessions could not be retrieved. The pipeline yielded 193 sets of contigs that probably belong to viruses. Similarity searches revealed that 34 sets of contigs had been previously published; thus, they were discarded. Out of the 159 sets remaining, only 81 passed the manual curation step (Fig. 9A).

Out of the 81 SRA accessions that contain curated genomes, 56 originated from plant samples (Fig. 9B), with 41 of those containing new species (Fig. 9C). However, only 26 plant accessions that contain putative new species originated from single individual samples (Fig. 9D). It is of notable interest that the overrepresentation of samples containing new species belonged to Gymnosperms. A single sample originating from non-vascular plants was sequenced from single branches and leaves of the livewort Mylia verrucosa [53].

Environmental samples constituted the second most significant source, contributing 18 accessions that resulted in novel genomes. Of these, 16 contained novel species. Specifically, 14 originated from soil samples, with three of them being derived from metagenomic studies investigating plant-associated microbiota. Curiously, one sample was collected in the sea, and the viral genome is probably the result of contamination.

Interestingly, three accessions came from arthropods: the grasshoppers Locusta migratoria and Ceracris nigricornis, and the phytophagous Apolygus lucorum. The latter is classified as an isolate of the Aspergillus flavus deltaflexivirus 1. The grasshopper Ceracris nigricornis was the only one that originated from a single individual (the Bioproject accession refers to an individual: PRJNA543568). The remaining four accessions originated from fungi: two from Macrocybe gigantea, an edible mushroom; two from the plant pathogen Puccinia striiformis.

Our results confirm the power of virus mining to provide insight into the complex landscape of viral diversity, which may lead to better taxonomic classifications. For instance, Emraviridae, a previously proposed family [46], was corroborated by the phylogenetic reconstruction, including our sequences and other similar genomes (Fig. 5). The sequences described in our work were clustered with genomes proposed for this family, Agaricus bisporus virus 5 and Agaricus bisporus virus 7. However, in the same work, the authors also suggest the creation of another family, Teagaviridae [46], but the results of this work do not support this separation. The phylogenetic reconstruction placed the viruses classified as Teagaviridae (Agaricus bisporus virus 7 and Agaricus bisporus virus 9) with members of the Deltaflexiviridae family (Fig. 6).

There is experimental evidence that the Sclerotinia sclerotiorum deltaflexivirus 1 (ScDV1 in Fig. 2) has a viral particle [23]. Additionally, there is also evidence for horizontal transmission in this family [54]. A gene annotated as a coat protein has orthologs present in members of the Deltaflexiviridae and the Thetaflexiviridae. Meanwhile, the gene annotated as HP3 has a distribution complementary to the coat protein gene. It is unclear whether HP3 codes for a coat protein, and this is a possible avenue for future research.

A well-known weakness of virus mining in public databases is the difficulty in establishing host-virus relationships, although inferences can be made in that regard [55]. The members of the Emraviridae possibly infect only fungi, since all the viruses in this group seem to be isolated from fungi, and the viruses described in this work originated from environmental samples [46]. Although it is not possible to ascertain the hosts of the viruses from the Deltaflexiviridae and Thetaflexiviridae, there is evidence pointing to both fungi [23, 54] and plants [5], albeit the latter was inferred indirectly.

Beyond the processing obstacles, uncertainty regarding the quality of the data deposited can also impede progress [12, 56]. The lack of metadata related to the SRA accessions precludes reaching many conclusions regarding virus ecology. It is of particular interest that there are annotations regarding the exact collection date and collection site. Only 36 of the curated viruses in this study had information on the collection date. Another potential obstacle is the classification of the discovered viruses. The common requirements usually involve wet lab experiments, however, most of the time, the samples are not available for such investigations [56]. However, collaborative efforts are taking place to open discussions on the directions of taxonomic classification in the post-genomics era [57, 58]

This work aimed to elucidate the diversity within the Deltaflexiviridae family. A striking diversity, both in terms of evolutionary history and genomic content, was found, supporting the creation of two new families and 33 new species. The new species proposed showed an identity below 90% to any other polyprotein sequence and were mined from samples in which a single individual was sampled (Supplementary Table 2). Our results highlight the importance of large-scale virus mining in public databases. Focusing on specific taxonomic groups or environmental settings can help enhance our understanding and characterization of viral diversity.

Aknowledgements

The authors would like to express their gratitude to the Coordination of Superior Level Staff Improvement (CAPES) for the financial support that made this work possible.

Data availability

The datasets and scripts used in this work are available upon reasonable request from the corresponding author.

Funding

Jhon Eric Alves Fernandes was the recipient of a PhD fellowship provided by the Coordination of Superior Level Staff Improvement (CAPES).

Competing Interests

The authors declare no competing interests

Author contributions

The study conception and design was performed by F.L.M, T.N., and J.E.A.F., Data collection and analysis were performed by J.E.A.F. The first draft of the manuscript was written by J.E.A.F., T.N., and, F.L.M. All authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Leinonen R, Sugawara H, Shumway M (2011) The Sequence Read Archive. Nucleic Acids Res 39:D19–D21. https://doi.org/10.1093/nar/gkq1019
Shi M, Lin X-D, Tian J-H, et al (2016) Redefining the invertebrate RNA virosphere. Nature 540:539–543. https://doi.org/10.1038/nature20167
Paez-Espino D, Eloe-Fadrosh EA, Pavlopoulos GA, et al (2016) Uncovering Earth’s virome. Nature 536:425–430. https://doi.org/10.1038/nature19094
Bejerman N, Dietzgen RG, Debat H (2021) Illuminating the Plant Rhabdovirus Landscape through Metatranscriptomics Data. Viruses 13:1304. https://doi.org/10.3390/v13071304
Bejerman N, Debat H (2022) Exploring the tymovirales landscape through metatranscriptomics data. Arch Virol 167:1785–1803. https://doi.org/10.1007/s00705-022-05493-9
Bejerman N, Dietzgen R, Debat H (2023) Novel Tri-Segmented Rhabdoviruses: A Data Mining Expedition Unveils the Cryptic Diversity of Cytorhabdoviruses. Viruses 15:2402. https://doi.org/10.3390/v15122402
Debat H, Garcia ML, Bejerman N (2023) Expanding the Repertoire of the Plant-Infecting Ophioviruses through Metatranscriptomics Data. Viruses 15:840. https://doi.org/10.3390/v15040840
Lee BD, Neri U, Roux S, et al (2023) Mining metatranscriptomes reveals a vast world of viroid-like circular RNAs. Cell 186:646-661.e4. https://doi.org/10.1016/j.cell.2022.12.039
de la Higuera I, Kasun GW, Torrance EL, et al (2020) Unveiling Crucivirus Diversity by Mining Metagenomic Data. mBio 11:e01410-20. https://doi.org/10.1128/mBio.01410-20
Mifsud JCO, Gallagher RV, Holmes EC, Geoghegan JL Transcriptome Mining Expands Knowledge of RNA Viruses across the Plant Kingdom. J Virol 96:e00260-22. https://doi.org/10.1128/jvi.00260-22
Bejerman N, Dietzgen RG, Debat H (2022) Unlocking the Hidden Genetic Diversity of Varicosaviruses, the Neglected Plant Rhabdoviruses. Pathogens 11:1127. https://doi.org/10.3390/pathogens11101127
Langmead B, Nellore A (2018) Cloud computing as a platform for genomic data analysis and collaboration. Nat Rev Genet 19:208–219. https://doi.org/10.1038/nrg.2017.113
Davis-Turak J, Courtney SM, Hazard ES, et al (2017) Genomics pipelines and data integration: challenges and opportunities in the research setting. Expert Rev Mol Diagn 17:225–237. https://doi.org/10.1080/14737159.2017.1282822
Roux S, Enault F, Hurwitz BL, Sullivan MB (2015) VirSorter: mining viral signal from microbial genomic data. PeerJ 3:. https://doi.org/10.7717/peerj.985
Zhao G, Wu G, Lim ES, et al (2017) VirusSeeker, a computational pipeline for virus discovery and virome composition analysis. Virology 503:21–30. https://doi.org/10.1016/j.virol.2017.01.005
Kieft K, Zhou Z, Anantharaman K (2020) VIBRANT: automated recovery, annotation and curation of microbial viruses, and evaluation of viral community function from genomic sequences. Microbiome 8:. https://doi.org/10.1186/s40168-020-00867-0
Phumiphanjarphak W, Aiewsakun P (2024) Entourage: all-in-one sequence analysis software for genome assembly, virus detection, virus discovery, and intrasample variation profiling. BMC Bioinformatics 25:222. https://doi.org/10.1186/s12859-024-05846-y
Li Y, Wang H, Nie K, et al (2016) VIP: an integrated pipeline for metagenomics of virus identification and discovery. Sci Rep 6:23774. https://doi.org/10.1038/srep23774
Di Tommaso P, Chatzou M, Floden EW, et al (2017) Nextflow enables reproducible computational workflows. Nat Biotechnol 35:316–319. https://doi.org/10.1038/nbt.3820
Mölder F, Jablonski KP, Letcher B, et al (2021) Sustainable data analysis with Snakemake. F1000Research 10:33. https://doi.org/10.12688/f1000research.29032.2
Ewels PA, Peltzer A, Fillinger S, et al (2020) The nf-core framework for community-curated bioinformatics pipelines. Nat Biotechnol 38:276–278. https://doi.org/10.1038/s41587-020-0439-x
Edgar RC, Taylor B, Lin V, et al (2022) Petabase-scale sequence alignment catalyses viral discovery. Nature 602:142–147. https://doi.org/10.1038/s41586-021-04332-2
Li K, Zheng D, Cheng J, et al (2016) Characterization of a novel Sclerotinia sclerotiorum RNA virus as the prototype of a new proposed family within the order Tymovirales. Virus Res 219:92–99. https://doi.org/10.1016/j.virusres.2015.11.019
Lefkowitz EJ, Dempsey DM, Hendrickson RC, et al (2018) Virus taxonomy: the database of the International Committee on Taxonomy of Viruses (ICTV). Nucleic Acids Res 46:D708–D717. https://doi.org/10.1093/nar/gkx932
Bolton MD, Thomma BPHJ, Nelson BD (2006) Sclerotinia sclerotiorum (Lib.) de Bary: biology and molecular traits of a cosmopolitan pathogen. Mol Plant Pathol 7:1–16. https://doi.org/10.1111/j.1364-3703.2005.00316.x
Jia J, Fu Y, Jiang D, et al (2021) Interannual dynamics, diversity and evolution of the virome in Sclerotinia sclerotiorum from a single crop field. Virus Evol 7:veab032. https://doi.org/10.1093/ve/veab032
Mu F, Xie J, Cheng S, et al (2017) Virome Characterization of a Collection of S. sclerotiorum from Australia. Front Microbiol 8:2540. https://doi.org/10.3389/fmicb.2017.02540
Thakur M, Buniello A, Brooksbank C, et al (2024) EMBL’s European Bioinformatics Institute (EMBL-EBI) in 2023. Nucleic Acids Res 52:D10–D17. https://doi.org/10.1093/nar/gkad1088
Sayers EW, Beck J, Bolton EE, et al (2023) Database resources of the National Center for Biotechnology Information. Nucleic Acids Res 52:D33–D43. https://doi.org/10.1093/nar/gkad1044
Meleshko D, Hajirasouliha I, Korobeynikov A (2021) coronaSPAdes: from biosynthetic gene clusters to RNA viral assemblies. Bioinformatics 38:1–8. https://doi.org/10.1093/bioinformatics/btab597
Li D, Luo R, Liu C-M, et al (2016) MEGAHIT v1.0: A fast and scalable metagenome assembler driven by advanced methodologies and community practices. Methods San Diego Calif 102:3–11. https://doi.org/10.1016/j.ymeth.2016.02.020
Camacho C, Coulouris G, Avagyan V, et al (2009) BLAST+: architecture and applications. BMC Bioinformatics 10:421. https://doi.org/10.1186/1471-2105-10-421
Bushnell B (2016) BBMap short read aligner
Kearse M, Moir R, Wilson A, et al (2012) Geneious Basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics 28:1647–9. https://doi.org/10.1093/bioinformatics/bts199
Li L, Stoeckert CJ, Roos DS (2003) OrthoMCL: Identification of Ortholog Groups for Eukaryotic Genomes. Genome Res 13:2178–2189. https://doi.org/10.1101/gr.1224503
Jones P, Binns D, Chang H-Y, et al (2014) InterProScan 5: genome-scale protein function classification. Bioinforma Oxf Engl 30:1236–1240. https://doi.org/10.1093/bioinformatics/btu031
Paysan-Lafosse T, Blum M, Chuguransky S, et al (2023) InterPro in 2022. Nucleic Acids Res 51:D418–D427. https://doi.org/10.1093/nar/gkac993
Muhire BM, Varsani A, Martin DP (2014) SDT: a virus classification tool based on pairwise sequence alignment and identity calculation. PloS One 9:e108277. https://doi.org/10.1371/journal.pone.0108277
Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32:1792–1797. https://doi.org/10.1093/nar/gkh340
Nguyen L-T, Schmidt HA, von Haeseler A, Minh BQ (2015) IQ-TREE: a fast and effective stochastic algorithm for estimating maximum-likelihood phylogenies. Mol Biol Evol 32:268–274. https://doi.org/10.1093/molbev/msu300
Katoh K, Standley DM (2013) MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol Biol Evol 30:772–780. https://doi.org/10.1093/molbev/mst010
Capella-Gutiérrez S, Silla-Martínez JM, Gabaldón T (2009) trimAl: a tool for automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics 25:1972–1973. https://doi.org/10.1093/bioinformatics/btp348
Letunic I, Bork P (2021) Interactive Tree Of Life (iTOL) v5: an online tool for phylogenetic tree display and annotation. Nucleic Acids Res 49:W293–W296. https://doi.org/10.1093/nar/gkab301
Sayers EW, Beck J, Bolton EE, et al (2024) Database resources of the National Center for Biotechnology Information in 2025. Nucleic Acids Res 53:D20–D29. https://doi.org/10.1093/nar/gkae979
Brister JR, Ako-Adjei D, Bao Y, Blinkova O (2015) NCBI viral genomes resource. Nucleic Acids Res 43:D571-577. https://doi.org/10.1093/nar/gku1207
Deakin G, Dobbs E, Bennett JM, et al (2017) Multiple viral infections in Agaricus bisporus - Characterisation of 18 unique RNA viruses and 8 ORFans identified by deep sequencing. Sci Rep 7:2469. https://doi.org/10.1038/s41598-017-01592-9
Warren RL, Keeling CI, Yuen MMS, et al (2015) Improved white spruce (Picea glauca) genome assemblies and annotation of large gene families of conifer terpenoid and phenolic defense metabolism. Plant J Cell Mol Biol 83:189–212. https://doi.org/10.1111/tpj.12886
Ma Y, Wang J, Hu Q, et al (2019) Ancient introgression drives adaptation to cooler and drier mountain habitats in a cypress species complex. Commun Biol 2:213. https://doi.org/10.1038/s42003-019-0445-z
Hubbard A, Lewis CM, Yoshida K, et al (2015) Field pathogenomics reveals the emergence of a diverse wheat yellow rust population. Genome Biol 16:23. https://doi.org/10.1186/s13059-015-0590-8
Sun Y, Abbott RJ, Lu Z, et al (2018) Reticulate evolution within a spruce (Picea) species complex revealed by population genomic analysis. Evol Int J Org Evol 72:2669–2681. https://doi.org/10.1111/evo.13624
Zhang C, Zhang T, Luebert F, et al (2020) Asterid Phylogenomics/Phylotranscriptomics Uncover Morphological Evolutionary Histories and Support Phylogenetic Placement for Numerous Whole-Genome Duplications. Mol Biol Evol 37:3188–3210. https://doi.org/10.1093/molbev/msaa160
Liu P-L, Zhang X, Mao J-F, et al (2020) The Tetracentron genome provides insight into the early evolution of eudicots and the formation of vessel elements. Genome Biol 21:291. https://doi.org/10.1186/s13059-020-02198-7
Dong S, Zhao C, Zhang S, et al (2019) The Amount of RNA Editing Sites in Liverwort Organellar Genes Is Correlated with GC Content and Nuclear PPR Protein Diversity. Genome Biol Evol 11:3233–3239. https://doi.org/10.1093/gbe/evz232
Hamid MR, Xie J, Wu S, et al (2018) A Novel Deltaflexivirus that Infects the Plant Fungal Pathogen, Sclerotinia sclerotiorum, Can Be Transmitted Among Host Vegetative Incompatible Strains. Viruses 10:295. https://doi.org/10.3390/v10060295
Roux S, Hallam SJ, Woyke T, Sullivan MB (2015) Viral dark matter and virus-host interactions resolved from publicly available microbial genomes. eLife 4:e08490. https://doi.org/10.7554/eLife.08490
Lauber C, Seitz S (2022) Opportunities and Challenges of Data-Driven Virus Discovery. Biomolecules 12:1073. https://doi.org/10.3390/biom12081073
Simmonds P, Adriaenssens EM, Zerbini FM, et al (2023) Four principles to establish a universal virus taxonomy. PLOS Biol 21:e3001922. https://doi.org/10.1371/journal.pbio.3001922
Roux S, Adriaenssens EM, Dutilh BE, et al (2019) Minimum Information about an Uncultivated Virus Genome (MIUViG). Nat Biotechnol 37:29–37. https://doi.org/10.1038/nbt.4306

SuppTab1250506.xlsx
Supplementary Table 1: SRA accessions used in this work. Each accession is identified by its unique Run ID. The information on the sample origin and BioProject was obtained through EUtilities. The information on the samples in which a single individual was sequenced was manually collected. The source article was included when available; otherwise, information was obtained from the sample description on the BioProject webpage.
SuppTab2250926.xlsx
Supplementary Table 2: Viruses mined in this work. Viruses classified in this work, with their respective proposed family. The Novel species column marks the sequences in which the polyprotein shares less than 90% identity with the closest publicly available BLAST hit. The Single Individual Sample column indicates sequences that were mined from samples in which a single individual was sequenced. Viruses that fulfill both criteria are highlighted in grey. The average sequencing depth was calculated using BBmap.

Download PDF

Editorial decision: Minor Revision
05 Dec, 2025
Reviewers agreed at journal
07 Nov, 2025
Reviewers invited by journal
31 Oct, 2025
Editor assigned by journal
01 Oct, 2025
First submitted to journal
30 Sep, 2025

You are reading this latest preprint version

Mining viruses in public databases unveils the diversity within the Deltaflexiviridae family

Status:

Version 1

Abstract

Figures

Introduction

Methods

Assembly pipeline

Manual curation and annotation

Results

Pipeline efficiency

Discussion

Declarations

Aknowledgements

Data availability

Funding

Competing Interests

Author contributions

References

Supplementary Files

Status:

Version 1