The Serratus Explorer ² was used to identify transcriptome datasets potentially containing zybavirus-like viral genome sequences (Edgar et al., 2022). The ORF1+2p fusion protein of ZbV-Z (GenBank accession ANN12897) was selected as the query. Default settings were used, with an alignment identity range of 45%–100% and a score range of 50 to 100. Datasets were subsequently filtered to include only those with an average read length of at least 100 nucleotides, a paired-end layout, and sequencing performed on the Illumina platform. This filtering process yielded 32 transcriptome datasets, which were then downloaded from the SRA at the NCBI.

Raw transcriptome reads were quality trimmed using Sickle (version 1.33) ³ with parameters “-q 30 -l 55.” Trimmed reads were assembled into contigs using SPAdes (version 4.0.0) in “rnaviral” mode ⁴ (Bushmanova et al., 2019). Assembled contigs were screened by BLASTX to identify those showing similarity to the ORF1+2p fusion proteins of ZbV-Z, EgZV1, EgZV2, and XIZV. When multiple zybavirus-like contigs were present within a single dataset, consensus sequences were generated using CAP3 (version date 02/10/15). Open reading frames (ORFs) were predicted for each contig using ORFfinder ⁵ .

Protein sequences were aligned using MAFFT (version 7.526) with the “--auto” parameter ⁶ (Nakamura et al., 2018). The resulting alignments were filtered using ClipKIT (version 2.4.1) ⁷ to retain only well-aligned, phylogenetically informative positions (Steenwyk et al., 2020). A maximum likelihood phylogenetic tree was constructed using IQ-TREE (version 3.0.0) with the “-mset WAG,LG,JTT” option ⁸ (Minh et al., 2020). Bootstrap support values were estimated from 1,000 replicates using the UFBoot2 method (parameter “-B 1000”). The final tree was visualized using MEGA (version 12.0.11) ⁹ (Kumar et al., 2024). Pairwise sequences identities were calculated using the Sequence Demarcation Tool (SDT) (version 1.3) ¹⁰ (Muhire et al., 2025).

To identify ribosomal RNA (rRNA) contigs of potential fungal origin, assembled transcriptome contigs were compared to small subunit rRNA sequences from the EUKARYOME database (version 1.9.4) ¹¹ (Tedersoo et al., 2024) using discontiguous megaBLAST. Contigs with an aligned length of at least 500 bp and nucleotide sequence identity of 90% or higher to fungal small subunit rRNA were retained as putative fungal rRNA contigs.

To identify candidate +1 PRF sites in the newly identified zybavirus-like genome sequences, UUU_CGN-like motifs were systematically searched within the overlapping regions of ORF1 and ORF2. Alignments of PRF sites were visualized using BOXSHADE (version 3.31) ¹² , and sequence logos were generated using WebLogo (version 3.7.9) ¹³ .

The genus Zybavirus currently includes one officially approved species, Z. bailii (ZbV-Z), and three unclassified viruses: EgZV1, EgZV2, and XIZV, as listed in the NCBI Taxonomy database (accessed 1 May 2025) (Depierreux et al., 2016; Chen et al., 2022; Teng et al., 2022; Zhang et al., 2022). To expand the known diversity of this genus, we analyzed transcriptome datasets identified using the Serratus Explorer (Edgar et al., 2022), followed by quality trimming, genome assembly, and similarity-based screening. Seven zybavirus-like genome sequences were recovered (Table 1).

Among these, two contigs (SRP327700-2 and SRP313929-1) shared 94.1% and 99.7% nucleotide identity, respectively, with the ZbV-Z reference genome. ZbV-Z was originally identified in the budding yeast Z. bailii (Radler et al., 1993; Schmitt and Neuhausen, 1994; Depierreux et al., 2016). The contig SRP327700-2 was derived from transcriptome data of fermented grains used in Baijiu production (Wei et al., 2023). Since the yeast Z. bailii participates in the fermentation of Baijiu and Maotai (Xu et al., 2017; Wei et al., 2023), it is likely that Z. bailii, or a closely related fungus serving as the viral host, was present in the sample. The second contig, SRP313929-1, originated from a Drosophila melanogaster transcriptome (Huang et al., 2021). While fruit flies are unlikely to serve as natural hosts for ZbV-Z, they may carry associated fungi. Supporting this, putative fungal rRNA contigs showing 93.6%–100% identity with the yeast Z. bailii rRNA sequences were detected in both datasets (see Supplementary Material S2 for putative fungal rRNA contigs identified across all transcriptome datasets analyzed in this study), suggesting that Z. bailii or related fungi may serve as hosts of these viruses.

Five additional contigs (DRP004880-1, SRP327700-1, SRP036135-1, SRP117133-1, and SRP262753-1) exhibited sequence similarity to known zybavirus genomes, suggesting that they represent novel zybavirus-like genomes (Figure 1). Two of these, DRP004880-1 and SRP327700-1, originated from transcriptome data of the yeast Z. bailii and fermented grains, respectively. These contigs shared 99.7% nucleotide identity and likely represent the same viral genome, herein designated Zygosaccharomyces bailii virus Z2 (ZbV-Z2). The RdRp sequences of ZbV-Z and ZbV-Z2 share 55% amino acid identity, supporting their classification as distinct viral genomes. DRP004880-1 was selected as the representative sequence. ZbV-Z2 encodes a 290-amino acid ORF1p and a 1017-amino acid ORF1+2p fusion protein. A putative +1 programmed ribosomal frameshifting (PRF) site, UUU_CGU, was identified in the overlapping region of ORF1 and ORF2.

The contig SRP036135-1 was recovered from a centipede (Cryptops hortensis) transcriptome (Fernandez et al., 2014). Given that all known zybavirus members are mycoviruses, this virus likely infects a fungus associated with the centipede. A fungal rRNA contig sharing 90.1% identity with a Kickxellales fungus was identified in the dataset. This viral genome was named Cryptops hortensis-associated virus Z1 (ChaV-Z1). The ChaV-Z1 RdRp shares 51% amino acid identity with ZbV-Z. The ChaV-Z1 genome sequence contains incomplete ORFs: ORF1 and ORF2 lack portions of the N-terminal and C-terminal amino acid residues, respectively. A putative +1 PRF site, UUU_CCU, was identified, which differs slightly from the canonical UUU_CGN motif.

The contig SRP117133-1 was derived from a Drosophila suzukii transcriptome (Medd et al., 2018). Multiple fungal rRNA contigs with 90.4%–100% identity to known fungal sequences were identified, suggesting a fungal origin for the contig. This viral genome was named Drosophila suzukii-associated virus Z1 (DsaV-Z1). Its RdRp shares 46% identity with that of ZbV-Z. The DsaV-Z1 genome sequence has a complete ORF1, while ORF2 lacks part of its C-terminal region. A putative +1 PRF site, UUU_CGA, was identified in the overlapping region of ORF1 and ORF2.

The final contig, SRP262753-1, originated from metatranscriptome data of salt marsh sediment collected from Sand Creek Marshes, Massachusetts, USA. This dataset contained 59 fungal rRNA contigs from diverse species, suggesting a fungal host for the contig. This viral genome, named Sand Creek Marshes virus Z1 (SCMV-Z1), encodes complete ORFs. The SCMV-Z1 RdRp shares 35% amino acid identity with EgZV1. Two potential +1 PRF motifs, UUU_CGU and UUU_CGC, were identified.

All genome sequences identified in this study were deposited in NCBI GenBank under accession numbers BK069872–BK069878 and are available in Supplementary Material S3. Sequencing depth plots of genomic contigs are presented in Supplementary Material S4.

To assess the phylogenetic relationships of the newly identified zybavirus-like genome sequences within the family Amalgaviridae, a maximum likelihood phylogenetic tree was constructed using RdRp (ORF2) amino acid sequences (Figure 2). BLASTP searches were performed using the RdRp sequences of ZbV-Z2, ChaV-Z1, DsaV-Z1, and SCMV-Z1 to identify closely related viruses in the NCBI database. A total of 42 representative viruses were selected, including members of the genera Amalgavirus and Unirnavirus, as well as unclassified viruses within the family Amalgaviridae.

Three of the four newly identified viral genomes (ZbV-Z2, ChaV-Z1, and DsaV-Z1) grouped with ZbV-Z and XIZV to form a monophyletic clade with strong bootstrap support (value of 100). This result supports the classification of ZbV-Z2, ChaV-Z1, and DsaV-Z1 as novel members of the genus Zybavirus.

The fourth viral genome, SCMV-Z1, clustered with EgZV1 and EgZV2, which are two previously proposed members of Zybavirus, in a separate clade distinct from the one containing the type species ZbV-Z. This clade also included five additional unassigned viruses: Beihai barnacle virus 14 (BBV14), Capillidibolus rugosus virus 1 (CaRuV1), Itsystermes virus (ItsV), Physcomitrium patens amalgavirus 1 (PpAV1), and Plasmopara viticola lesion-associated amalga-like virus 1 (PvLaAV1) (Shi et al., 2016; Lay et al., 2020; Vendrell-Mir et al., 2021). This distinct lineage, referred to here as the “BBV14 clade” based on the earliest reported member, may represent a novel genus within the family Amalgaviridae. The bootstrap support value for this clade was 98.

The proposed genus “Anlovirus” was represented in the tree by AnloV1 and a closely related virus, HplV59 (Shi et al., 2016; Pyle et al., 2017; Teng et al., 2022; Zhang et al., 2022). Three additional unclassified viruses, Gutsystermes virus (GutV), Sanya amalgavirus 1 (SaAmV1), and XiangYun partiti-picobirna-like virus 7 (XYpplV7), also clustered within this lineage. Although SaAmV1 is currently annotated as a member of the genus Amalgavirus, and the other two viruses are listed as unclassified, their phylogenetic placement suggests that they may be more appropriately assigned to “Anlovirus.” The bootstrap support value for the “Anlovirus” clade, which includes five viruses, was 89.

The remaining viruses formed two well-supported clades corresponding to the genera Amalgavirus and Unirnavirus, each with a bootstrap support value of 100. These results are consistent with their current taxonomic classification.

To supplement the phylogenetic analysis, pairwise sequence identities among RdRp amino acid sequences were calculated (Supplementary Material S5). Pairwise identities ranged from 19.9% to 73.0%, with a mean of 29.8%. The sequences were clustered into known genera and proposed clades, consistent with the phylogenetic analysis. The mean pairwise RdRp sequence identities within distinct clades were as follows: Zybavirus, 47.5% (range: 42.8%–54.3%); “BBV14 clade,” 30.4% (range: 24.1%–41.0%); “Anlovirus,” 29.6% (range: 23.9%–43.4%); Amalgavirus, 48.6% (range: 41.9%–68.0%); and Unirnavirus, 58.5% (range: 47.6%–73.0%). The “BBV14 clade” and “Anlovirus” exhibited greater diversity than the genera Zybavirus, Amalgavirus, and Unirnavirus, suggesting that the former two clades may comprise highly divergent viruses that could be classified into multiple independent genera.

The UUU_CGN sequence has been proposed as a slippery consensus motif that induces +1 PRF during translation of the ORF1+2p fusion protein in members of the genus Amalgavirus (Nibert et al., 2016). This motif was also identified within the overlapping region of ORF1 and ORF2 in three of the newly identified viral genomes in this study: ZbV-Z2, DsaV-Z1, and SCMV-Z1. In contrast, the viral genome of ChaV-Z1 contained a UUU_CCU sequence, which differs slightly from the canonical UUU_CGN motif.

To evaluate the conservation of +1 PRF motifs, we analyzed 18 viral genomes classified in this study as members of the genus Zybavirus (five viruses), the “BBV14 clade” (eight viruses), and the proposed genus “Anlovirus” (five viruses). Among the five viruses assigned to Zybavirus, three (DsaV-Z1, ZbV-Z, and ZbV-Z2) possessed the UUU_CGN motif. The remaining two, ChaV-Z1 and XIZV, contained UUU_CCU and UUU_CUU motifs, respectively (Figure 3A). Among the eight viruses grouped in the “BBV14 clade,” six (CaRuV1, EgZV1, EgZV2, PpAV1, PvLaAV1, and SCMV-Z1) contained the canonical UUU_CGN motif. The other two, BBV14 and ItsV, contained UUU_CUU and UUU_CUA sequences, respectively (Figure 3B). Among the five viruses classified as members of the proposed genus “Anlovirus,” only GutV contained a UUU_CGN motif. The remaining viruses included AnloV1 and HplV59 with UUU_CUU, SaAmV1 with UUU_CAA, and XYpplV7 with UUU_CCG (Figure 3C).

Overall, ten of the 18 analyzed genomes contained the UUU_CGN motif. The remaining eight genomes retained a UUU codon followed by a C nucleotide. These findings suggest that UUU_CNN may represent a broader consensus +1 PRF motif for this group of viruses.

For comparison, we also examined the +1 PRF motifs in members of the genus Amalgavirus. Most species retained the UUU_CGN motif, as previously reported (Figure 3D), although some deviations have been noted in earlier studies (Nibert et al., 2016).

In contrast, members of the genus Unirnavirus are thought to employ a −1 PRF mechanism. A conserved slippery sequence, G_GAU_UUU, has been proposed to mediate this translation event (Campo et al., 2016; Depierreux et al., 2016). Among the 13 ICTV-approved species of Unirnavirus, ten contained the G_GAU_UUU motif, while two contained G_GAU_UUA and one had G_GAU_UUC (Figure 3E). Although G_GAU_UUU is the most frequently observed motif, the broader consensus sequence G_GAU_UUN may better encompass the diversity within this genus.