Viral diversity and ecological functions in the world’s largest deep-sea mineral exploration region

Bowen Hou; Lilan Zhang; Dong Sun

doi:10.1038/s41467-026-78161-0

. 2026 Sep 29;17:11227. doi: 10.1038/s41467-026-78161-0

Viral diversity and ecological functions in the world’s largest deep-sea mineral exploration region

Bowen Hou ^1,², Lilan Zhang ^1,^3,^✉, Dong Sun ^2,^4,^5,^✉

PMCID: PMC13720004 PMID: 42957726

Abstract

Deep-sea polymetallic nodule provinces host distinct benthic prokaryotic communities, yet viruses that regulate them remain largely unknown. Here, we established the viral genome dataset of the Clarion-Clipperton Fracture Zone, the world’s largest deep-sea mineral exploration region, to investigate viral diversity and ecological functions. Protein-sharing network and phylogenetic analyses showed extensive benthic viral diversity, with over 99% of viral clusters not represented in current reference databases. Inferred virus–host associations suggest that viral infections span a broad range of prokaryotic lineages involved in carbon, nitrogen, phosphorus, sulfur, and metal cycling. By infecting key functional microbes and encoding auxiliary metabolic genes, viruses may modulate host metabolism and survival during infection, with potential implications for biogeochemical processes in polymetallic nodule field sediments. Moreover, gene exchange analyses indicate the potential of viruses to facilitate host adaptation to metal stress. These findings provide insights into the ecological roles of viruses in the polymetallic nodule province, establishing a vital pre-mining baseline of microbial ecosystem for future environmental impact assessments.

Subject terms: Ecology, Microbial biooceanography, Evolutionary ecology, Marine biology, Microbial ecology

The viral contributions to deep-sea polymetallic ecosystems, rich in valuable mineral nodules, remain unclear. Here, the authors created the first large-scale catalog of viruses from the Clarion-Clipperton Fracture Zone and found extensive viral diversity that may influence nutrient cycling, microbial survival, and adaptation to metal-rich environments.

Introduction

Deep-sea benthic ecosystems harbor distinct and abundant microbial communities and genetic resources that play crucial roles in global biogeochemical cycles^1,2. Among them, the deep-sea polymetallic nodule province represents a distinct abyssal habitat characterized by sediments enriched with polymetallic nodules that contain abundant iron, manganese, and associated trace metals, which are widely distributed across the abyssal plains of the tropical Pacific and Indian Oceans³. The formation of polymetallic nodules is controlled by hydrogenetic precipitation of metal oxides from seawater and diagenetic processes of metals driven by microbial degradation of organic matter within sediments⁴. The mosaic of soft sediments and hard polymetallic nodules generates highly heterogeneous microscale redox environments⁵, exposing local microbial habitats to persistent metal ion stress. Moreover, the absence of terrestrial inputs from continents and the oligotrophic conditions in the upper ocean result in low sedimentation rates (<1 mm kyr⁻¹) and organic carbon supply, making this habitat one of the most oligotrophic environments in the deep ocean^6,7. Long-term oxidative-diagenetic processes have produced distinct vertical geochemical gradients in the sediments, where concentrations of organic carbon, dissolved metals, and nitrate change over scales ranging from a few to tens of centimeters with sediment depth⁸. These geochemical conditions are expected to harbor distinct species and genetic resources.

To survive in this oligotrophic, metal-rich habitat, microorganisms inevitably evolve metabolic adaptations and metal-tolerance mechanisms. Functional profiling of dominant prokaryotes in sediments of polymetallic nodule provinces has found that these microorganisms mainly have the potential to utilize inorganic nutrients through redox reactions rather than conventional organic nutrient metabolism to sustain survival⁹. In addition, some taxa exhibit remarkable capabilities for metal resistance and transformation^9,10, and are involved in Mn and Fe oxidation processes, thereby potentially contributing to the nucleation and growth of polymetallic nodules^6,11,12. Viruses have been increasingly recognized as key mediators of host regulation and gene exchange in marine microbial communities^13,14. These interactions have been proposed to contribute to virus–host coevolution in marine microbial communities¹⁵. Similar processes may also occur in polymetallic nodule field sediments, potentially contributing to the formation of distinct microbial assemblages. However, viral diversity and evolutionary relationships in this habitat remain largely unexplored.

Abundant benthic viral diversity and functional signatures have recently been observed in the sediments of cold seeps, hydrothermal vents, trenches, and seamounts^16–19. As estimated, through infection and lysis, viruses release cellular contents that contribute substantially to benthic biogeochemical processes, accounting for about 35% of total prokaryotic metabolism in deep-sea sediments and serving as a major source of labile organic matter, regenerated nutrients, and bioavailable metals^13,20. Beyond recycling processes, viruses can directly modulate host metabolism through auxiliary metabolic genes (AMGs)^21,22. These genes are involved not only in energy acquisition, carbon fixation, and nutrient transformation, but also in metal redox reactions, transport, and resistance^23–25, potentially influencing the biogeochemical processes associated with polymetallic nodule formation. In addition, viruses can promote host evolution and adaptation in metal-rich environments by mediating horizontal gene transfer^26,27. In the deep-sea polymetallic nodule provinces, viruses may serve as key agents of coevolutionary adaptation, modulating prokaryotic hosts, sustaining microbial activity, and driving biogeochemical cycles within these metal-rich, energy-limited environments, yet these ecological functions remain unproven and largely unexplored.

The Clarion-Clipperton Fracture Zone (CCZ) in the eastern Pacific Ocean represents the world’s largest polymetallic nodule province, containing at least 21 billion tons of dry nodules and designated as a primary target area for future deep-sea mining⁴. In this study, we collected sediment profile samples for metagenomic sequencing from 14 stations across the CCZ and integrated them with previously published metagenomic datasets from the same region⁹. Then, a comprehensive sediment viral genome dataset for the CCZ was established to reveal the viral diversity, virus–host interactions, and AMGs related to biogeochemical cycling. This study advances our understanding of viral ecology in this deep-sea ecosystem and provides baseline data for evaluating the ecological impacts of future deep-sea mining from the perspective of virus–microbe interactions.

Results

A total of 53 sediment samples spanning four depth intervals (0–2, 4–6, 8–10, and 14–16 cm below the seafloor, cmbsf) were collected from 14 stations in the CCZ polymetallic nodule province of the eastern tropical Pacific Ocean (Fig. 1 and Supplementary Data 1). Each sample was subjected to both metagenomic and 16S rRNA gene sequencing. In addition, seven previously published metagenomic samples (0–5 cmbsf) from this region were retrieved from the Sequence Read Archive (SRA) at NCBI⁹. A total of 1.2 Tb of high-quality metagenomic data was obtained after combining and filtering the datasets described above. Subsequent analyses were conducted to recover viral and prokaryotic genomes from all 60 samples and to examine the spatial patterns of viral and prokaryotic communities using the 53 depth-profile samples collected in this study. Corresponding prokaryotic community patterns are provided in the Supplementary Results and Supplementary Figs. 1, 2.

Fig. 1 — The inset map shows that the sampling sites are located in the eastern tropical Pacific Ocean. Map created using the Free and Open Source QGIS v3.40.7¹¹⁸ with the GEBCO_2025 Grid¹¹⁹. Red points indicate sampling sites from this study, and purple points indicate metagenomic datasets obtained from the Sequence Read Archive (SRA) at NCBI. Source data are provided as a Source Data file.

Establishment of the Clarion-Clipperton fracture zone virome (CCZV)

Based on five viral identification approaches, 14,181 viral sequences longer than 5 kb were recovered from contigs generated by de novo assembly of the metagenomes. These viral sequences were clustered at 95% average nucleotide identity (ANI), resulting in 11,742 viral operational taxonomic units (vOTUs), representing species-level taxa²⁸, hereafter referred to as the Clarion-Clipperton Fracture Zone Virome (CCZV; Supplementary Data 2). The completeness of vOTUs in the CCZV ranged from short fragments to complete genomes, including 519 complete, 279 high-quality, and 769 medium-quality genomes (Supplementary Fig. 3a). Genomic sizes ranged from 5 to 493.6 kb, with 4919 vOTUs longer than 10 kb. The largest mean genome size was observed for high-quality genomes (54.2 kb), followed by complete (52.0 kb), medium-quality (41.4 kb), low-quality (10.8 kb), and not-determined genomes (10.8 kb). Notably, 16 vOTUs exceeded 200 kb in length, suggesting that they may represent putative jumbo phages²⁹. Virus-encoded proteins were clustered at 60% identity and 80% coverage, yielding 242,084 viral protein clusters (vPCs).

According to the viral taxonomy proposed by the International Committee on Taxonomy of Viruses (ICTV), 95.7% of the vOTUs received taxonomic assignments (Fig. 2a). The vast majority of vOTUs (94.8%) belonged to the class Caudoviricetes within the phylum Uroviricota (Supplementary Fig. 3b), which dominate most environments, such as marine, terrestrial, and human gut ecosystems^18,30–32. However, 95.6% of the vOTUs could not be assigned to any known family (Fig. 2a). Genome protein-sharing network analysis showed that only 16 viral clusters (VCs; 0.9%) from the CCZV were shared with the NCBI Viral RefSeq (Supplementary Fig. 4a, b and Supplementary Data 3), whereas 99.3% of vOTUs (n = 11,662) were not clustered with any reference sequences, indicating limited representation of CCZV viruses in current reference datasets.

Given the dominance of Caudoviricetes in the CCZ and other environments, a phylogenetic tree was constructed for medium- to high-quality CCZV vOTUs assigned to Caudoviricetes. The tree was inferred from a concatenated alignment of 77 marker proteins³³ and included reference Caudoviricetes genomes from NCBI Viral RefSeq (Supplementary Fig. 5). A total of 5384 viral genomes, including 840 CCZV vOTUs and 4544 reference sequences, were subjected to multiple sequence alignment. Many viruses formed distinct clades closely related to known families, such as Chimalliviridae, Zobellviridae, Mesyanzhinovviridae, Herelleviridae, and Peduoviridae. Some of the identified jumbo viruses clustered within the Chimalliviridae clade, a recently proposed viral family that typically comprises jumbo viruses with genomes exceeding 200 kb and encodes multiple anti-host defense genes, and whose members have attracted interest for phage therapy³⁴.

In addition, based on the presence of lysogenic marker proteins in viral genomes, 1,038 vOTUs (8.8%) were inferred to be temperate viruses (Fig. 2b). This estimate was further supported by an analysis restricted to highly complete vOTUs (Supplementary Fig. 3c). The lifestyles of the remaining vOTUs were conservatively classified as unknown, as they may include diverse non-lysogenic strategies such as lytic, stable carrier, or chronic infection states.

Depth-related patterns in viral communities, populations, and protein clusters

Principal component analysis (PCA) showed partial separation among sediment layers for viral community composition, although samples still exhibited substantial overlap (Fig. 2c). This pattern was supported by permutational multivariate analysis of variance (PERMANOVA), which detected significant differences among depth layers (pseudo-F = 4.46, R² = 0.28, P = 0.001), and a similar trend was observed for viral functional composition (Supplementary Fig. 6a). Nucleotide diversity (π) of vOTUs showed a significant decreasing trend across sediment depth layers (Fig. 2d). Approximately 90% of viral genes exhibited pN/pS < 1, and median pN/pS values (~0.28–0.31) were similar across depth layers (Fig. 2e), indicating that purifying selection predominated throughout the sediment profile. Viral genes with pN/pS > 1 accounted for 9.9–10.7% of genes across sediment depth layers. These genes were mainly associated with DNA replication, repair, and metabolism, as well as virion assembly and packaging (Fig. 2f), suggesting that adaptive evolution or selection pressure may primarily target functions linked to genome maintenance and virion assembly. Additional genome-level trait comparisons showed modest variation among sediment depth layers, with largely overlapping distributions despite statistically significant differences (Supplementary Fig. 7).

Depth-related patterns of vOTUs and vPCs were further examined, with detailed results provided in the Supplementary Results. A total of 723 vOTUs (8%) were shared across all sediment depth layers (Supplementary Fig. 8a), and 117 dominant vOTUs accounted for 44.5% of total viral abundance (Supplementary Fig. 8b). Depth-related vOTUs identified using DESeq2 were grouped into four trend groups, with the 0–2 cmbsf-associated and 14–16 cmbsf-associated groups showing opposing abundance trends (Supplementary Fig. 9). At the vPC level, layer-specific vPCs ranged from 9371 to 57,268, corresponding to 19.3–52.9% of detected vPCs within each layer. Overall, 36.6% of vPCs were shared among multiple layers, including 6.6% shared across all layers (Supplementary Fig. 10). Functional annotation showed broadly similar high-level categories across layers (Supplementary Fig. 11).

A total of 10 environmental variables were measured in this study, including Fe, Mn, Cu, Ni, Zn, Se, Mo, Si, total organic carbon (TOC), and total nitrogen (TN; Supplementary Data 1; their distributions are described in the Supplementary Results and shown in Supplementary Figs. 12–14). Mantel tests showed that viral community composition was significantly correlated with Zn, Se, Ni, and Mn in the 0–2 cmbsf layer and with Zn and Cu in the 14–16 cmbsf layer (P < 0.05), whereas no significant correlations were detected in the intermediate layers (Supplementary Fig. 15a). These results suggest that detectable virus–environment correlations were layer-specific.

Overall, these results indicate that viral communities, populations, and protein clusters in CCZ sediments exhibited depth-associated variation across the sediment profile. However, the presence of shared vOTUs and vPCs across sediment layers, together with broadly similar high-level functional categories among annotated vPCs, suggests that sediment layers retained a degree of similarity in viral population and functional composition.

Cross-ecosystem comparison reveals extensive heterogeneity in CCZV viral lineages and proteins

To investigate the relationships between CCZV viruses and those from other marine environments, we integrated vOTU datasets from seawater (GOV 2.0)³⁰, trenches (OTVGB)³⁵, cold seeps¹⁷, and hydrothermal vents³⁶ with the CCZV for joint analysis. All vOTUs with sequence lengths ≥10 kb were analyzed using vCONTACT2 to construct a viral genome protein-sharing network, resulting in 17,104 VCs representing genus-level viral groups³⁷. Among these, GOV 2.0, CCZV, hydrothermal vents, trenches, and cold seeps contributed 14,494, 1065, 1571, 1060, and 845 VCs, respectively (Fig. 3a and Supplementary Data 4 and 5). Only 8 VCs were shared across all marine environments (Fig. 3b), indicating that strong habitat heterogeneity drives viral endemism and highlighting the necessity for broader investigation and exploration. The CCZV contained 2,418 vOTUs clustered into 1,065 VCs. Of these, 580 VCs were shared with other marine environments (Supplementary Fig. 16), with the largest number shared with GOV 2.0 (n = 360), followed by trenches (n = 252), hydrothermal vents (n = 117), and cold seeps (n = 63). The remaining 485 VCs were detected only in the CCZV among the currently compared reference datasets (Fig. 3b).

Fig. 3 — a Gene-sharing network of viral sequences (≥10 kb) across different marine environments, including sediments of deep-sea polymetallic nodule provinces, seawater, cold seep sediments, hydrothermal vent sediments, and trench sediments. Nodes represent viral sequences, and edges connect significantly similar viral sequences. Node colors denote the origin environment of viral sequences. b Venn diagram showing the number of shared viral clusters between the CCZV and other marine environmental viral genome datasets. c The UpSet plot summarizes overlap in viral protein clusters (vPCs) between the CCZV and viral datasets from hydrothermal vents, trenches, and cold seeps. Connected intersection points indicate vPCs shared by the CCZV and the corresponding ecosystem(s). The gray bars indicate the number of shared vPCs in each intersection, with the values labeled above each bar. The accompanying donut charts show the functional composition of viral proteins contained within those shared vPCs. Bars on the left depict the total number of protein clusters in each dataset (gray) and the subset shared with CCZV (colored). Source data are provided as a Source Data file.

Functional annotation of viral proteins from the CCZV and other deep-sea datasets, including hydrothermal vent, cold seep, and trench datasets, showed that most proteins were either unannotated or assigned to unknown functions, with the highest proportion observed in the CCZV (84.0%; Supplementary Fig. 17). The high-level functional composition of annotated viral proteins was broadly similar across datasets. To further assess habitat specificity and the extent of protein sharing, viral proteins from the CCZV and three other deep-sea datasets were clustered. Across the four datasets, 581,022 protein sequences were recovered and clustered into 463,858 vPCs. The CCZV, hydrothermal vent, trench, and cold seep datasets contained 163,800, 177,404, 76,603, and 55,056 vPCs, respectively (Fig. 3c). Only 6,562 CCZV vPCs (4.0%) were shared with at least one other deep-sea dataset. The greatest overlap was observed with the trench dataset (5,666 vPCs), whereas overlap with hydrothermal vent and cold seep datasets was much smaller (674 and 529 vPCs, respectively; Fig. 3c), consistent with the pattern observed for VCs (Supplementary Fig. 16). Among the vPCs shared between CCZV and other deep-sea datasets, annotation rates (24.0–51.9%) were higher than for the full CCZV set (16%), and annotated shared vPCs were enriched in conserved functions related to virion structure and assembly. Notably, only 37 vPCs were shared across all four datasets (Fig. 3c), spanning 7 functional categories and indicating a small set of broadly conserved viral proteins across deep-sea habitats. Overall, these results suggest limited protein sharing between the CCZV and other deep-sea viral datasets, despite broad similarity in the high-level functional composition of annotated proteins.

Virus–host linkages in CCZ sediments

To predict virus–host linkages in CCZ sediments, 2,345 medium- to high-quality prokaryotic metagenome-assembled genomes (MAGs) were recovered from de novo assembly and binning of 60 metagenomic samples. After dereplication at 95% ANI, 484 species-level MAGs were obtained (Supplementary Data 6), spanning 31 bacterial and 2 archaeal phyla. Among bacteria, most MAGs were assigned to Pseudomonadota (n = 137), Acidobacteriota (n = 51), Planctomycetota (n = 48), Bacteroidota (n = 35), and Chloroflexota (n = 32). Archaeal MAGs were assigned to Thermoproteota (n = 19) and Nanoarchaeota (n = 6). Notably, 78.7% of MAGs remained unclassified at the species level. Functional profiling of the 484 MAGs indicated the community was dominated by heterotrophic life strategies (Supplementary Fig. 18 and Supplementary Data 7), with the potential for organic carbon oxidation detected in nearly all genomes (Supplementary Fig. 19). In addition, metabolic potentials linked to Fe and selected metalloid redox transformations were also detected in multiple taxa (Supplementary Fig. 18). Metal resistance genes (MRGs) were broadly distributed across CCZ MAGs, with cross-ecosystem (cold seep, hydrothermal vent, and trench datasets) comparisons showing a higher prevalence of MRG-carrying genomes but lower MRG density among MRG-carrying genomes than in several other deep-sea datasets (Supplementary Fig. 20).

A total of 755 virus–host linkages were identified using the iPHoP pipeline based on CCZ MAGs, involving 482 vOTUs (4.1% of 11,742) and 192 MAGs (Supplementary Data 7). Phylogenetic analysis linked the vOTUs to a broad range of hosts spanning 26 bacterial and 2 archaeal phyla. Only 5 phyla among the MAGs showed no detectable viral associations (Fig. 4a), indicating the near-universal presence of viral infection across prokaryotic lineages. Approximately 2.8% of these viruses were assigned to known families, whereas the majority remained within unclassified Caudoviricetes (Fig. 4b). At the host phylum level, Planctomycetota, Pseudomonadota, Thermoproteota, Acidobacteriota, Bacteroidota, and Nitrospirota were the most frequently identified potential hosts. These hosts encoded genes associated with carbon, nitrogen, phosphorus, sulfur, and metal cycling, including genes related to iron and manganese transformations (Supplementary Fig. 21 and Supplementary Data 8). These observations suggest that viral predation may potentially affect microbially mediated biogeochemical cycles.

Fig. 4 — a Maximum-likelihood phylogenetic trees of the recovered bacterial (blue) and archaeal (brown) metagenome-assembled genomes (MAGs) at the phylum level were inferred from a concatenated alignment of 120 bacterial or 53 archaeal single-copy marker genes. The visualization style was inspired by Yu et al.¹⁸ and Gios et al.¹²⁰. This figure was generated using original data from this study. Both articles are published under the Creative Commons Attribution 4.0 International License (CC BY 4.0; https://creativecommons.org/licenses/by/4.0/). The orange circles indicate the number of vOTUs predicted to have a host in the clade. b Predicted virus-host linkages between viral taxonomy (at family level) and recovered bacterial and archaeal MAGs in the CCZV. c Relative abundances of vOTUs and their prokaryotic hosts grouped by the host taxonomy in each sediment samples. Top bars represent samples from different depth layers. Source data are provided as a Source Data file.

Viruses predicted to infect specific host taxa showed abundance patterns broadly similar to those of their corresponding microbial hosts (Fig. 4c and Supplementary Fig. 22a), consistent with observations from other habitats such as cold seeps, groundwater, and permafrost^17,31,38. This concordance indicates tight ecological coupling between viruses and their hosts, as further supported by a strong regression between host and lineage-specific viral abundance (R² = 0.79, P < 0.001; Supplementary Fig. 22b). To further explore their associations, we next examined one-to-one correlations between the relative abundance of each virus and its predicted host. Temperate viruses generally showed weaker abundance correlations with their predicted hosts than other viral groups (Supplementary Fig. 23). The range of virus-to-host relative abundance ratios (VHR) was broadly consistent across depth layers (Supplementary Fig. 22c), suggesting relatively stable infection dynamics along the vertical profile. At the phylum level, most computed VHR values were greater than 1 (Supplementary Fig. 22d), which can be interpreted as reflecting active viral production arising from either lytic infections or non-lytic strategies involving continuous virus release. Among these, Thermoproteota and Pseudomonadota exhibited the highest relative abundances, and viruses predicted to infect these phyla constituted the most abundant fraction of the viral community. This pattern is consistent with the kill-the-winner framework, in which high population densities of dominant microorganisms increase encounter rates with their specific viruses³⁹.

Virus-carried metal resistance genes

A total of 40 vPCs representing 16 metal-resistance categories (8 multi-metal and 8 single-metal) were identified as MRGs in CCZV (Supplementary Fig. 24a and Supplementary Data 9). Among them, 12 genes were annotated as multi-metal MRGs and 28 as single-metal MRGs. Single-metal MRGs were primarily associated with resistance to Ni, Zn, Hg, and Cu, and were detected throughout the sediment profile. Notably, Ni-resistance genes consistently represented the most abundant single-metal resistance category across all depth layers. Further analyses showed that both multi-metal MRGs and several single-metal MRG categories were significantly correlated with environmental variables in different depth layers (Supplementary Fig. 25). Overall, these associations were predominantly positive in the 0–2 cmbsf layer but tended to become negative in the 14–16 cmbsf layer.

To assess whether the MRGs detected on vOTUs might reflect virus–microbe genetic exchange, we compared MRG homologs between vOTUs and microbial MAGs based on the reciprocal best BLAST hit approach. Notably, 15 potential gene exchange events involving Ni, Zn, Cu, Hg, Fe, and multi-metal MRGs were identified between viruses and microbial MAGs, suggesting that viruses may participate in the dissemination of MRGs within the microbial community (Supplementary Data 9). For instance, three MRGs in A5_MC06a_2.bin.7 (nikB and nikC for Ni tolerance; zraR/hydH for Zn tolerance/regulation) showed high similarity to sequences in two vOTUs, implying possible virus-mediated gene transfer. These findings suggest that viruses may promote dissemination of MRGs within the microbial community, potentially enhancing host metal-tolerant capabilities^25,40.

Potential roles of viral AMGs in biogenic element cycles

Viruses can supplement or reprogram host metabolism through the expression of AMGs, thereby participating in biogeochemical processes^14,41. To investigate the potential ecological roles of viruses in the CCZ, the DRAM-v pipeline was applied and results were manually curated. From 4,920 vOTUs ≥ 10 kb, a total of 293 putative AMGs were identified (Supplementary Data 10). Of these, 52 putative AMGs were assigned to 5 KEGG metabolism categories based on their KO terms, including Metabolism of cofactors and vitamins (n = 18), Energy metabolism (n = 13), Carbohydrate metabolism (n = 12), Amino acid metabolism (n = 7), and Xenobiotics biodegradation and metabolism (n = 2; Supplementary Fig. 26). A subset of putative AMGs was detected across all depth layers and in most samples, including energy metabolism–related genes (metF, mer, nuoL, and sdhB) and amino acid metabolism–related genes (dapB and hcaC; Fig. 5a). In contrast, some putative AMGs showed depth preferences. For example, PGLS (6-phosphogluconolactonase), a gene encoding a key step in the oxidative branch of the pentose phosphate pathway and associated with oxidative stress responses, was detected more frequently in the 0–2 cmbsf layer (Fig. 5a).

Putative AMGs were linked to functions relevant to carbon, nitrogen, sulfur, and phosphorus metabolism (Fig. 5b). Among these, AMGs related to carbon metabolism were especially prominent. Under oligotrophic and energy-limited conditions, microbial survival may depend on metabolic flexibility and the maintenance of cellular homeostasis⁴². Within this framework, the pentose phosphate pathway (PPP) may play an important role in supporting redox balance and biosynthetic demand. Notably, two PPP-related AMGs, PGLS (n = 1) and TALDO1 (n = 4), were identified in CCZV viruses. Additional carbon metabolism-related AMGs were also detected, including genes associated with glycolysis/gluconeogenesis (PK, n = 2; eno, n = 1; PGI, n = 1; pgi-pmi, n = 1) and one-carbon metabolism (metF, n = 2; mer, n = 1). For sulfur metabolism, cysH (n = 1) related to assimilatory sulfate reduction was detected. For nitrogen metabolism, gdhA (n = 1) associated with ammonium assimilation was detected. For phosphorus metabolism, phoH (n = 10) related to the phosphate-starvation response was detected. Gene-context analysis showed that these genes were flanked by viral-like and hallmark genes, supporting a viral origin. Moreover, TALDO1, eno, mer, phoH, and cysH were detected in temperate viruses, with some occurring in genomes that also encoded phage integrases (Supplementary Fig. 27), suggesting that these AMGs may contribute to host metabolism during lysogenic association.

Potential involvement of viruses in metal cycling

Interestingly, six putative AMGs encoding heme oxygenase (HO) homologs were identified in CCZV viral genomes, and these genes were detected in 5 samples (8.3% of all samples). Genome-context inspection indicated that these putative AMGs were located within viral genomic regions and were flanked by vviral-likeand hallmark genes, supporting a viral origin (Fig. 6a). HO catalyzes oxidative cleavage of heme to biliverdin and related tetrapyrroles, with concomitant release of Fe(II) and carbon monoxide⁴³. Although the viral HO homologs showed low amino acid sequence identity to previously reported microbial HO proteins (<40%), structural prediction revealed substantial similarity in their tertiary structures (Fig. 6b and Supplementary Fig. 28). Phylogenetic analysis showed that these HO homologs were most closely related to Myxococcota-associated HemO-like sequences (Supplementary Fig. 29). Reports of virus-encoded HO genes associated with HemO-like homologs remain limited. By contrast, cyanophage HO homologs have been experimentally linked to bilin metabolism during infection⁴⁴. Collectively, these observations support the possibility of virus-encoded HO homologs as putative AMGs in CCZ sediments and suggest potential impacts on host heme and Fe turnover during infection. Further functional evidence will be required to evaluate ecological consequences.

Fig. 6 — a Genome map of six virus-encoded heme oxygenase genes (HO1–HO6). b Tertiary structures of selected AMGs based on structural modeling using Colabfold. The color gradient represents the predicted local distance difference test (pIDDT) score, where higher scores indicate higher confidence. c Maximum-likelihood phylogenetic trees of the recovered bacterial and archaeal MAGs. Phylogenetic trees of were inferred from a concatenated alignment of 120 bacterial or 53 archaeal single-copy marker genes, and visualized using the online platform iTOL⁷⁵. Source data are provided as a Source Data file.

Furthermore, viruses may indirectly influence metal cycling by modulating the metabolism of putative metal-cycling hosts through AMGs. Specifically, among the 23 host MAGs linked to viruses carrying AMGs, 21 exhibited potential for Fe/Mn redox metabolism, including 6 MAGs encoding Fe-oxidation genes (Fig. 6c and Supplementary Data 11). Of these 6 MAGs, 5 contained cytochrome–porin (Cyc2), spanning Pseudomonadota, Planctomycetota, and KBS1 lineages. Cyc2 encodes an outer-membrane cytochrome–porin that is proposed to oxidize extracellular Fe(II) and funnel electrons into the cellular electron transport chain, thereby supporting iron oxidation-linked energy metabolism⁴⁵. For example, one vOTU linked to DY88_A5_MC03_4.bin.3 encoded TALDO1, suggesting that viruses may indirectly support putative metal-cycling hosts by redistributing pentose phosphate-associated carbon flux and potentially contributing to intracellular redox balance during infection⁴⁶. Overall, our results suggest that viruses may participate in metal cycling in ferromanganese nodule sediments by directly encoding metal cycling-related AMGs and indirectly modulating the metabolic activities of putative metal-cycling hosts.

Discussion

This study established the CCZV using 60 metagenomic sediment samples collected from a deep-sea polymetallic nodule field. Genome protein-sharing network analysis suggested that CCZV viruses are distinct from viral genomes reported in previously studied environments and those deposited in the NCBI viral reference database (Fig. 3a and Supplementary Fig. 4). Cross-ecosystem comparisons showed limited overlap of shared VCs and viral proteins between CCZV and other deep-sea datasets (Fig. 3), suggesting that these viruses may possess distinct genes to adapt to deep-sea polymetallic nodule fields. Notably, more VCs and vPCs were shared between the CCZV and trench datasets than between the CCZV and cold seep or hydrothermal vent datasets. These patterns are consistent with differences in habitat energy supply. Both CCZ sediments and many trench environments are heterotrophic and rely largely on allochthonous organic-matter inputs^6,47, whereas cold seeps and hydrothermal vents are fueled by reduced chemical compounds and typically support strong in situ chemoautotrophic production^48,49. Together, these observations suggest that environmental filtering related to energy availability may contribute to shaping viral community structure across distinct deep-sea ecosystems.

Mantel tests revealed associations between viral community composition and selected environmental variables in CCZ sediments (Supplementary Fig. 15a), suggesting that local environmental conditions may be related to viral community structure. As one of the world’s largest deep-sea polymetallic nodule provinces, the CCZ is characterized by two major environmental constraints: long-term metal enrichment and oligotrophy^4,50. In this environmental context, CCZV viruses exhibited corresponding adaptive features.

For adaptation to metal-rich conditions, CCZV viruses may be involved in both host metal detoxification and metal-associated biogeochemical processes. A higher prevalence of MRG-carrying genomes was observed in CCZ MAGs than in other representative deep-sea habitats (Supplementary Fig. 20), consistent with sustained selection on host resistance functions under long-term metal enrichment^9,51. Virus-encoded MRGs were also detected in CCZV genomes, including genes associated with multi-metal resistance as well as resistance to Ni, Zn, Hg, and As (Supplementary Fig. 24a). In surface samples, some of these genes were positively correlated with Ni, Zn, and Cu concentrations (Supplementary Fig. 25). Because elevated exposure to these metals can disrupt intracellular metal homeostasis and induce oxidative stress, they may constitute a persistent selective background for virus–host interactions^52,53. Potential MRG exchange between viruses and MAGs further suggests that viruses may facilitate the movement of these resistance traits within local communities (Supplementary Data 9). This is consistent with previous studies proposing that viruses can act as mobile reservoirs of resistance traits in environmentally stressed systems^27,54. Taken together, the virus-encoded MRGs detected in CCZ sediments may represent a potential mechanism by which resistance functions are maintained and circulated under long-term metal selection. Evidence for a viral role in metal cycling was also identified. Six putative AMGs encoding HO were identified in viral genomes (Fig. 6a). HO catalyzes heme cleavage and releases Fe(II)⁴³, which may increase the bioavailability of heme-bound iron and reduce heme-associated oxidative stress^55,56. Under oxic conditions, any Fe(II) that escapes cellular sequestration can be rapidly oxidized to Fe(III) and precipitate as Fe oxides⁵⁷. Together, these results suggest that CCZV viruses may contribute both to metal stress responses and to metal-associated biogeochemical processes.

For adaptation to oligotrophic conditions, CCZV viruses may be involved in host carbon metabolic processes. Microbial metabolism in CCZ sediments is sustained largely by photosynthetically derived organic matter exported from the euphotic zone^50,58,59. Both in situ evidence⁵⁸ and functional annotation indicate that energy acquisition in this system is closely linked to organic matter degradation (Supplementary Figs. 18 and 19), whereas viruses may further act on key metabolic nodes in this process through AMGs (Fig. 5b). Virus-encoded PGLS suggests potential viral modulation of host carbon metabolism. The gene encodes 6-phosphogluconolactonase, which functions in intermediate turnover in the oxidative PPP and helps maintain flux through this branch, thereby indirectly supporting NADPH production⁶⁰. NADPH plays a central role in antioxidant defense and cellular redox homeostasis. Notably, virus-encoded PGLS was mainly detected in the surface layer (Fig. 5a), consistent with the higher concentrations of Ni, Cu, and Zn observed there. This further highlights viral adaptation to metal-rich conditions through carbon metabolism-associated functions. In addition, metF and mer were detected in nearly all samples and at relatively high abundance (Fig. 5a), suggesting broad viral involvement in host one-carbon metabolic processes. Given the persistently oligotrophic and low-energy conditions of CCZ sediments⁵⁰, efficient recycling and reuse of limited substrates may be particularly important⁶¹. In this context, virus-encoded metF and mer are more likely to reflect regulation of host one-carbon metabolic networks and may help maintain metabolic homeostasis during infection. Taken together, these findings suggest that CCZV viruses may target key metabolic nodes in this oligotrophic, metal-rich habitat and thereby contribute to host adaptation under extreme environmental conditions.

Although depth-associated differences revealed vertical heterogeneity in viral communities across the sediment profile, several shared viral and functional features were also observed across the sediment profile, including core viral populations present across all layers, relatively conserved high-level functional categories, and broadly similar virus–host interaction patterns (Fig. 4c and Supplementary Figs. 11, 22). These shared features may provide a basis for community reassembly after disturbance. This is particularly relevant for assessing the ecological consequences of deep-sea mining, because polymetallic nodule fields are major targets for future deep-sea mineral resource development⁶². Mining activities will induce physical disturbance, resuspension, and redeposition of surface sediments⁶³, and long-term disturbance experiments have shown that the effects of such disturbance on nodule-field sediments can persist for decades⁶⁴. Under this scenario, viral populations and functional modules shared between deeper and surface sediments could provide a potential source for partial reassembly of surface viral communities after disturbance. However, this inference remains to be tested through in situ mining-disturbance observations and controlled simulation experiments.

In conclusion, this study reveals extensive viral diversity in polymetallic nodule sediments and highlights the potential involvement of viruses in host adaptation to metal-enriched, oligotrophic conditions and in microbially mediated elemental cycling within this habitat. Altogether, this study provides a critical baseline for assessing microbial ecosystem functions prior to deep-sea mining and for informing the conservation and management of deep-sea ecosystems. Nevertheless, the ecological roles inferred here are based primarily on genome-resolved and comparative analyses, and many proposed functions remain to be validated experimentally. Future work integrating virome sequencing, transcriptomics, cultivation-based infection assays, and disturbance simulations will be essential for linking viral genomic potential to ecological function and for predicting how deep-sea viral communities respond to mining-related sediment disturbance.

Methods

Sample collection and element analysis

The Clarion–Clipperton Zone (CCZ) in the eastern Pacific hosts the world’s largest polymetallic nodule province, spanning ~5.9 × 10⁶ km² and containing an estimated ~21.1 billion dry tons of nodules⁴. As one of the most important areas under the jurisdiction of the International Seabed Authority for future deep-sea mineral development, the CCZ currently contains 17 of the 31 ISA exploration contracts⁸. Sediment disturbance is widely recognized as one of the major environmental consequences of deep-sea mining, and recovery of microbially mediated functions may require decades⁶⁴. A total of 14 sediment sampling stations were therefore established in a polymetallic nodule field in the western CCZ (9–11°N, 154–157°W) (Supplementary Data 1), where sediment samples from 0–16 cmbsf were collected to characterize viral communities in shallow nodule-field sediments, thereby providing foundational knowledge and baseline data for future assessments of mining-related disturbance. Sediment samples were collected by a multiple corer sampler (the inner diameter of the sampling tube is 10 cm) during the COMRA DY79 (September-November 2023) and DY88 (August-October 2024) cruises. The sediment samples were sliced into 2 cm per layer from the surface to 16 cm on board and frozen at −80 °C until DNA extraction in laboratory. After quality assessment, a total of 53 samples were finally used in this study.

The sediment samples were freeze-dried, homogenized, and finely ground. Cu, Ni, Zn, and Mo were measured by inductively coupled plasma mass spectrometry (ICP-MS) after microwave digestion with a mixture of HNO₃, HF, and H₂O₂, while Se was determined by hydride generation atomic fluorescence spectrometry after digestion. Fe and Mn contents were determined by wavelength-dispersive X-ray fluorescence spectrometry. For TOC and TN measurements, sediment samples were dried, acidified with HCl to remove carbonates, washed to neutrality, dried again, and homogenized. Subsamples were then encapsulated in tin capsules and analyzed using an elemental analyzer (Elementar Analysensysteme, Germany).

DNA extraction, 16S rRNA gene sequencing and analysis

Total DNA was extracted from 10 g of sediment per sample using the DNeasy PowerMax Soil Kit (Qiagen, Germany) according to the manufacturer’s instructions. The purity and quantity of extracted DNA were checked using Quantus™ Fluorometer (Promega, Madison, WI, USA). The hypervariable region V3-V4 of the bacterial 16S rRNA gene was amplified with primer pairs 338 F (5′-ACTCCTACGGGAGGCAGCAG-3′) and 806 R (5′-GGACTACHVGGGTWTCTAAT-3′) by T100 Thermal Cycler PCR thermocycler (BIO-RAD, USA). The PCR reaction mixture contained 4 μL 5 × Fast Pfu buffer, 2 μL 2.5 mM dNTPs, 0.8 μL each primer (5 μM), 0.4 μL Fast Pfu polymerase, 10 ng of template DNA, and ddH2O to a final volume of 20 µL. PCR amplification cycling conditions were as follows: initial denaturation at 95 °C for 3 min, followed by 27 cycles of denaturing at 95 °C for 30 s, annealing at 55 °C for 30 s and extension at 72 °C for 45 s, and single extension at 72 °C for 10 min, and end at 4 °C. The PCR product was extracted from 2% agarose gel and purified using the PCR Clean-Up Kit (YuHua, Shanghai, China) according to the manufacturer’s instructions and quantified using Qubit 4.0 (Thermo Fisher Scientific, USA). Purified amplicons were pooled in equimolar amounts and paired-end sequenced on an Illumina Nextseq 2000 platform (Illumina, San Diego, USA) according to the standard protocols by Majorbio Bio-Pharm Technology Co. Ltd. (Shanghai, China).

After demultiplexing, the resulting sequences were quality-filtered with fastp v0.19.6 and merged with FLASH v1.2.11⁶⁵. Then the high-quality sequences were de-noised using DADA2 plugin in the Qiime2 v2024 pipeline with recommended parameters⁶⁶, which obtains single nucleotide resolution based on error profiles within samples. DADA2 denoised sequences are usually called amplicon sequence variants (ASVs). Taxonomic classification of ASVs was performed using the q2-feature-classifier classify-sklearn (Naive Bayes) method implemented in QIIME 2, against the SILVA v138.2 reference database, with a confidence threshold of 70%.

Metagenomic sequencing and assembly

DNA extracts were used for library preparation with the NEXTFLEX Rapid DNA-Seq Kit (Bioo Scientific, USA) and sequenced on a NovaSeq X Plus (Illumina, San Diego, California, USA) by Majorbio Bio-Pharm Technology Co., Ltd. (Shanghai, China). The raw reads were trimmed and quality filtered using fastp v0.22.0⁶⁷ (parameters: -q 30 -n 5 -w 2 -l 50) to obtain clean reads with high quality. Clean reads in each sample were de novo assembled using MEGAHIT v1.2.9⁶⁸ with default parameters, and then remove contigs less than 1000 bp. Additionally, seven publicly available metagenomic datasets from CCZ sediment samples (Supplementary Data 1) were retrieved from the NCBI SRA database and processed following the same quality control and assembly procedures described above.

Reconstruction of prokaryotic metagenome-assembled genomes

MAGs were recovered using MetaWRAP v1.2.1⁶⁹ with default parameters and three binning methods (CONCOCT v0.4.0, MaxBin v2.2.2 and MetaBAT v2.12.1). The resulting MAGs were assessed for MAG completeness and contamination using CheckM v1.0.12⁷⁰ (parameters: -c 50 -x 10). Based on the Minimum Information about a Metagenome-Assembled Genome (MIMAG) criteria⁷¹, medium-quality-level MAGs with ≥50% completeness and ≤10% contamination were retained for downstream analysis. Then, all genomes were de-replicated at 95% average nucleotide identity (ANI) using dRep v1.4.3⁷² (parameters: -comp 50 -conn 10 -sa 0.95), resulting in a total of 484 species-level MAGs. Taxonomy of MAGs was determined using GTDB-Tk v2.3.2 with the GTDB database r214⁷³. Phylogenetic trees of MAGs were calculated using FastTree v2.1.11⁷⁴ (parameters: -gamma -wag) based on the 120 bacterial or 53 archaeal universal marker genes, and visualized using the online platform iTOL⁷⁵. Functional annotation of MAGs was performed using the “METABOLIC-C.pl” module in METABOLIC v4.0⁷⁶ (parameters: -m-cutoff 0.75 -kofam-db full).

Identification of viral contigs

To recover as many viral sequences as possible, assembled contigs were analyzed using five different viral identification methods, namely, JGI pipeline^41,77, VIBRANT v1.2.1⁷⁸, Virsorter2 v2.2.4⁷⁹, DeepVirfinder v1.0⁸⁰ and PPR-meta v1.1⁸¹. VIBRANT, Virsorter2 and PPR-meta use default parameters, and sequences identified by DeepVirfinder with a predictive score threshold above 0.85 and a significance value below 0.05 were putative viral contigs. JGI pipeline used the default three filters, which were five hits to viral protein families combined with other criteria. Putative viral contigs from the five identification methods described above were subjected to host contamination removal and completeness estimation using CheckV v1.0.1⁸² with default parameters, and contigs annotated as no viral gene detected and shorter than 5 kb were further filtered. The identified viral contigs (n = 14,181) were clustered based on a 95% ANI and a minimum of 85% alignment coverage, following the standards defined by the Minimum Information about an Uncultivated Virus Genome (MIUViG)²⁸. Specifically, we performed an all-vs-all BLAST search v2.13.0+⁸³ (parameters: -task megablast -perc_identity 90 -max_target_seqs 25,000). The ANI and alignment fraction (AF) between sequence pairs were calculated using a Python script (https://bitbucket.org/berkeleylab/checkv/src/master/scripts). Based on these metrics, the identified viral sequences were clustered into viral operational taxonomic units (vOTUs; n = 11,742).

Lifestyle prediction

The life strategies of vOTUs were inferred using two complementary pipelines. (1) VIBRANT v1.2.1⁷⁸ and CheckV v1.0.1⁸² were employed to identify vOTUs with provirus integration signals or integrase genes, which are classified as temperate viruses. (2) Twenty-nine lysogeny-related HMM profiles (transposase, integrase, excisionase, resolvase, and recombinase) were retrieved from Pfam database as previously described²⁶, and searched against viral proteins using HMMER v3.3.2⁸⁴ with the “-cut_ga” flag. vOTUs showing significant hits to these HMMs were considered temperate viruses⁸⁵. The vOTUs identified by either of the above approaches were classified as temperate, while others were designated as unknown.

Clustering of viral proteins and taxonomic annotation of vOTUs

Viral genes were predicted using prodigal v2.6.3⁸⁶ (parameters: -p meta). All proteins encoded by viruses were clustered using CD-HIT v4.8.1⁸⁷ (parameters: -c 0.6 -aS 0.8 -n 4 -T 0 -g 1) to generate viral protein clusters (vPCs) to represent viral functional repertoires. To determine the taxonomic affiliations of the vOTUs, taxonomic annotation was conducted using geNomad v1.7.6 with default parameters⁸⁸. Viral genes within the vOTUs were annotated based on taxonomically informative marker profiles in geNomad, and the vOTUs were subsequently classified into viral lineages according to the Virus Metadata Resource of the ICTV.

Abundance profiles

RPKM (Reads per kilobase per million mapped reads) values were used to represent relative abundances of vOTUs, vPCs and MAGs. For vOTUs and vPCs, clean reads of each sample were mapped to representative sequences using BWA-MEM v0.7.17-r1188⁸⁹ with default parameters. The resulting SAM files were converted to sorted BAM files using SAMtools v1.17⁹⁰ with default parameters and subsequently processed with CoverM v0.6.1 (https://github.com/wwood/CoverM) in contig mode to filter low-quality mappings and generate RPKM and count profiles for each sample (parameters: --trim-min 0.10 --trim-max 0.90 --min-read-percent-identity 0.95 --min-read-aligned-percent 0.90 -m rpkm, count). Only vOTUs with a covered fraction ≥75% were retained for downstream analyses. For MAGs, the same steps and parameters as those used for vOTUs and vPCs were applied, with CoverM executed in genome mode.

Gene characteristic calculations for GC content, protein molecular weight, and protein elemental composition

Following previous methods⁹¹, viral genomic sequences and he amino acid sequences translated from their predicted genes were input into the Python script get_gc_and_narsc.py (available at https://github.com/faylward/pangenomics/) to calculate GC content, molecular weight, and C/N/S-ARSC values.

Prediction of virus–host linkages

Prior to constructing the custom MAG database, all viral sequences were searched against all microbial MAGs using BLASTN v2.13.0+⁸³ to identify potential viral contamination arising from misbinning. MAG contigs showing 100% nucleotide identity and 100% alignment coverage to viral sequences were considered viral contaminants and were removed to minimize false-positive host assignments based on sequence homology. Following this decontamination step, the remaining medium- to high-quality MAGs were added to the iPHoP database using the “add_to_db” option. Host prediction was then performed using iPHoP v1.4.1⁹² with default parameters. Only predictions pointing to CCZ MAGs were retained.

Functional annotation and AMG identification

Viral proteins were functionally annotated using VIBRANT v1.2.1⁷⁸ default parameters, based on hmmsearch against the KEGG, VOG, and Pfam HMM databases. For each protein, the best-supported annotation was selected according to the highest bit score, followed by the lowest e-value and then the highest viral score. Functional categories were further standardized using a custom Python script (https://github.com/cody-mar10/protein_set_transformer/blob/main/manuscript/protein_embeddings/annotations/relabel_VOG.ipynb).

Putative AMGs encoded by vOTUs (≥10 kb) were identified and annotation using DRAM-v v1.5.0⁹³ pipeline with default parameters. vOTU sequences were first processed using VirSorter2 v2.2.4⁷⁹ with the “–prep-for-dramv” option, and putative AMGs were subsequently predicted from the processed sequences by DRAM-v with an auxiliary score threshold of <4. To minimize false positives and to follow a conservative interpretation framework for viral AMGs⁹⁴, putative AMGs were manually curated by (i) screening functional annotations to remove genes more likely to represent viral core or infection-related functions, and (ii) inspecting the genomic context of each candidate on the viral contig. Specifically, genes assigned to nucleotide metabolism, DNA replication/repair–related processes, modification of viral components, ribosomal proteins, transcriptional/translational regulators, and viral invasion/host interaction were removed. Only genes that are flanked by viral-like or viral hallmark genes and are not at the immediate edge of the contig were retained as putative AMGs. Three-dimensional structures of putative AMGs were predicted using ColabFold v1.5.2⁹⁵, which is based on AlphaFold2, and the resulting models were visualized in ChimeraX⁹⁶. Selected protein structural homology searches and predictions were performed using the Phyre2 web portal⁹⁷. The relative abundance of putative AMGs is determined by calculating the relative abundance of vOTUs carrying those AMGs in each sample.

Comparisons of viral genomes to public databases

To compare the viral sequences from the CCZV with those from other marine environments, protein-sharing network analysis of CCZ sediment vOTUs, Global Oceans Viromes 2 (GOV 2.0) database (≥10 kb, n = 195,728)³⁰, and viral contigs from hydrothermal vents (n = 7552), cold seeps (n = 2885), and trenches (n = 3817) were performed. For each viral contig, open reading frames (ORFs) were predicted using Prodigal v2.6.3⁸⁶ (parameters: -p meta -g 11 -m-c), and predicted protein sequences were then subjected to all-to-all BLASTp using Diamond v2.0.15⁹⁸; the result file was used as input for vConTACT2 v0.11.3³⁷ with default parameters. Viral RefSeq (v211) was used as reference database to generate the protein-sharing network, and Cytoscape v3.9.1⁹⁹ was used to visualize the network.

To compare viral proteins from CCZV with those from other deep-sea ecosystems, viral sequences ≥10 kb from CCZV, trenches, hydrothermal vents, and cold seeps were annotated using the same VIBRANT-based workflow described above. Viral proteins from the four habitats were clustered with CD-HIT v4.8.1⁸⁷ (parameters: -c 0.6 -aS 0.8 -n 4 -T 0 -g 1) to generate viral protein clusters, which were subsequently used to quantify shared and habitat-specific clusters across habitats.

Microdiversity analyses of viral genomes

The MetaPop pipeline was executed using the default parameters (--snp_scale both) to calculate viral microdiversity¹⁰⁰. To ensure accurate single-nucleotide polymorphism (SNP) calling and reliable contig-level microdiversity estimation, only viral populations with >10× average read depth and >70% genome coverage were retained for analysis. SNP frequencies were subsampled to 10× coverage to calculate π at both the gene and genome levels, as well as selective pressures on specific genes (pN/pS). For each sample, π values of 50 viral genomes were randomly selected and averaged. This was repeated 1000x and the average of all 1000 subsamplings was used as the final microdiversity value for each sample.

Phylogenetic analyses

A concatenated protein phylogeny of Caudoviricetes genomes was constructed as previously described^31,33. Seventy-seven marker proteins were identified from vOTUs over medium-quality and NCBI RefSeq viral genomes using HMMER v3.3.2⁸⁴. Specifically, HMMs for the 77 markers were searched against protein sequences, and the best hits were selected. Each marker alignment was individually trimmed using trimAl v1.5.rev0¹⁰¹ (parameter: -gt 0.5), and the alignments were concatenated by filling gap positions for missing markers. Only genomes containing at least three markers were retained. Genomes with <5% alignment columns were subsequently removed. The phylogeny of Caudoviricetes genomes was inferred from the concatenated multiple sequence alignment using FastTree v2.1.11⁷⁴ under the WAG + G model. The midpoint-rooted tree was visualized with iTOL, and family -level taxonomic annotations for the NCBI RefSeq viral genomes were obtained directly from the Virus Metadata Resource of ICTV.

Sequences similar to AMGs were recruited from the NCBI nr database using BLASTp v2.13.0+⁸³ searches against the viral AMGs (bit score > 100; E-value < 1 × 10⁻⁵). Each set of viral AMGs were then aligned with Muscle v5.1¹⁰² with default parameters and filtered by trimAl v1.5.rev0¹⁰¹ to remove columns comprised of more than 95% gaps. Maximum-likelihood phylogenetic trees were then inferred using FastTree v2.1.11⁷⁴ with default parameters and visualized in iTOL.

Metal resistance gene identification

To compare the prevalence of MRG-carrying MAGs between CCZ MAGs and those from other deep-sea ecosystems, dereplicated MAGs from this study, trenches^103,104, hydrothermal vents⁴⁷, and cold seeps¹⁷ were included. MAG datasets from other deep-sea ecosystems were separately dereplicated at 95% ANI using dRep v1.4.3⁷² (parameters: -comp 50 -conn 10 -sa 0.95) prior to comparison. All MAGs were then searched against the BacMet v2.2 database¹⁰⁵ using BLASTp v2.13.0+⁸³. Hits were retained at ≥60% identity, ≥80% coverage, and an E-value < 1 × 10⁻⁵. Cross-ecosystem comparisons were further performed within different MAG quality categories, defined by completeness ranges of 50–70%, 70–90%, and ≥90%.

Annotation of virus-encoded MRGs was performed using BLASTp searches against the BacMet v2.2 database¹⁰⁵. Hits were retained at ≥ 40% identity, ≥ 80% coverage, and an E-value < 1 × 10⁻⁵.

Gene exchange analysis

Potential gene exchanges between viruses and curated MAGs were inferred using the reciprocal best BLAST hit approach¹⁰⁶. Briefly, virus-encoded MRGs were searched against gene sets from MAGs using BLASTp v2.13.0+⁸³ (E-value ≤ 1 × 10⁻⁵) to identify their best cellular homologs. Subsequently, a combined BLASTp database containing both viral and MAG-derived genes was constructed, and the identified cellular homologs were queried against this database. A gene pair was considered a putative virus–prokaryote gene exchange if the best cellular homolog retrieved a top hit to the same virus-encoded gene as in the initial BLASTp search, excluding self-hits to the same cellular homolog.

Statistical analyses and plotting

All statistical analyses were performed in R v4.3.1¹⁰⁷ with a significance level of P < 0.05 unless otherwise mentioned. Alpha diversity was calculated using the R packages vegan v2.6-10¹⁰⁸ and picante v1.8.2¹⁰⁹. Statistical differences among groups were evaluated using the Kruskal-Wallis and Wilcoxon rank-sum tests implemented in the stats package¹⁰⁷. Mantel tests were performed using the vegan package to assess correlations between dissimilarity matrices. Environmental variables were log₁₀(x + 1)-transformed and z-score standardized prior to ordination analyses performed using the R packages vegan v2.6-10¹⁰⁸. Linear mixed-effects models were fitted using the R packages lme4 v1.1-37¹¹⁰, with site included as a random effect to account for spatial heterogeneity among sampling sites.

To investigate vertical differences in prokaryotic and viral community structure across sediment depths, count data were analyzed using a CLR transformation to handle the inherent compositional nature of the high-throughput sequencing data¹¹¹. Prokaryotic abundances were derived from read counts obtained by 16S rRNA gene amplicon sequencing, whereas vOTUs and vPCs abundances were based on read counts generated from read mapping using CoverM. Because the CLR transformation is undefined in the presence of zero values, zeros were first replaced using the “cmultRepl” function (method = “CZM”) implemented in the zCompositions R package¹¹². The resulting data were then CLR-transformed using the “decostand” function in the vegan package. PCA was performed separately for each dataset on the CLR-transformed abundance data using the “prcomp” function in R (equivalent to PCA in Aitchison distances). Differences in community composition among sediment depth layers were assessed separately for each dataset using PERMANOVA, implemented with the “adonis2” function in the vegan package and likewise based on Aitchison distances.

Following previously research¹¹³, differential abundance analysis was performed on vOTU read count data using DESeq2 in R project¹¹⁴. To reduce the multiple-testing burden introduced by low-abundance features, a pre-filtering step was applied prior to modeling, retaining only vOTUs with at least 20 counts in at least 5 samples. Count data were modeled in DESeq2 using negative binomial generalized linear models, with sediment depth as the main explanatory factor and sampling site included to account for spatial heterogeneity. The overall effect of depth was evaluated using a likelihood ratio test. P values were adjusted for multiple testing using the false-discovery rate method, and vOTUs with adjusted P values < 0.05 were defined as depth differential vOTUs.

To characterize depth-associated enrichment patterns of differentially abundant vOTUs, RPKM values were used for trend summarization. Within each sample, RPKM values of the differential vOTUs were converted to relative abundances by normalizing to the total RPKM of all differential vOTUs in that sample. For each vOTU, relative abundances were standardized across samples using z-score transformation, and mean z-scores were then calculated within each depth group to represent depth-associated enrichment profiles. Euclidean distances based on these mean profiles were used for hierarchical clustering with Ward’s D2 method, and the resulting dendrogram was cut into four clusters. Cluster depth preference was assigned using a majority-vote approach based on the depth at which each vOTU reached its highest mean z-score, and the four resulting clusters were designated as the 0–2, 4–6, 8–10, and 14–16 cmbsf-associated trend groups. Depth-associated trend groups were visualized using faceted boxplots of mean z-scores across sediment depths.

Stacked bar plots, box plots, donut charts, and scatter plots (e.g., Figs. 1, 4c) were generated using ggplot2 v3.5.2¹¹⁵, heat maps (e.g., Fig. 5a and Supplementary Fig. 12) using pheatmap v1.0.12 (https://CRAN.R-project.org/package=pheatmap), gene maps (e.g., Fig. 6a) using gggenes v0.5.1 (https://CRAN.R-project.org/package=gggenes), Venn diagrams (e.g., Supplementary Fig. 14) using VennDiagram v1.7.3¹¹⁶, UpSet plots (e.g., Fig. 3c and Supplementary Fig. 18) using UpSetR v1.4.0¹¹⁷, and Sankey diagrams (e.g., Fig. 4b) using networkD3 v0.4.1 (https://CRAN.R-project.org/package=networkD3).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Supplementary Information^{(29.8MB, pdf)}

41467_2026_78161_MOESM2_ESM.pdf^{(6.1KB, pdf)}

Description of Additional Supplementary Files

Supplementary Data 1-11^{(16.4MB, xlsx)}

Reporting Summary^{(3.3MB, pdf)}

Transparent Peer Review file^{(35.5MB, pdf)}

Source data

Source Data^{(4.4MB, xlsx)}

Acknowledgements

We are grateful to the members of R/V DAYANGYIHAO for their help during the investigation.

Author contributions

L.L.Z. and D.S. designed the research. B.W.H. conducted the bioinformatic and statistical analysis. B.W.H., L.L.Z., and D.S. wrote the paper. All authors read and approved the final manuscript.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This work was funded by the National Natural Science Foundation of China (42522712, U25A6027, 42377387), National Key Research and Development Program of China (2022YFC2806805) and the Digital Deep-sea Typical Habitats Program of China Deep Ocean Affairs Administration.

Data availability

The sequencing reads generated in this study have been deposited in the CNCB database under accession code PRJCA049271. The sequencing data used in this study are publicly available in the NCBI database under accession code PRJNA808646. The representative viral genomes, microbial genomes, and processed data generated in this study are available in the Zenodo repository at https://doi.org/10.5281/zenodo.20629959. Source data are provided with this paper.

Code availability

The custom code used for viral identification and figure generation are available at https://github.com/Bowencqu/CCZV/ or https://doi.org/10.5281/zenodo.21306299.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Lilan Zhang, Email: lilanzhang@cqu.edu.cn.

Dong Sun, Email: sund@sio.org.cn.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-78161-0.

References

1.Chen, J. et al. Global marine microbial diversity and its potential in bioprospecting. Nature633, 371–379 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Zhang, C., Peng, Y., Liu, X., Wang, J. & Dong, X. Deep-sea microbial genetic resources: new frontiers for bioprospecting. Trends Microbiol32, 321–324 (2024). [DOI] [PubMed] [Google Scholar]
3.Hein, J. R., Mizell, K., Koschinsky, A. & Conrad, T. A. Deep-ocean mineral deposits as a source of critical metals for high- and green-technology applications: comparison with land-based resources. Ore Geol. Rev.51, 1–14 (2013). [Google Scholar]
4.Koschinsky, A. & Hein, J. R. Marine Ferromanganese Encrustations: Archives of Changing Oceans. Elements13, 177–182 (2017). [Google Scholar]
5.Simon-Lledó, E. et al. Ecology of a polymetallic nodule occurrence gradient: Implications for deep-sea mining. Limnol. Oceanogr.64, 1883–1894 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Shiraishi, F. et al. Dense microbial community on a ferromanganese nodule from the ultra-oligotrophic South Pacific Gyre: Implications for biogeochemical cycles. Earth Planet. Sci. Lett.447, 10–20 (2016). [Google Scholar]
7.Versteegh, G. J. M., Koschinsky, A., Kuhn, T., Preuss, I. & Kasten, S. Geochemical consequences of oxygen diffusion from the oceanic crust into overlying sediments and its significance for biogeochemical cycles based on sediments of the northeast Pacific. Biogeosciences18, 4965–4984 (2021). [Google Scholar]
8.China Minmetals Corporation. Environmental Impact Statement (China Minmetals Corporation, 2025).
9.Zhang, D. et al. Microbe-driven elemental cycling enables microbial adaptation to deep-sea ferromanganese nodule sediment fields. Microbiome11, 160 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Gillard, B., Chatzievangelou, D., Thomsen, L. & Ullrich, M. S. Heavy-Metal-Resistant Microorganisms in Deep-Sea Sediments Disturbed by Mining Activity: An Application Toward the Development of Experimental in vitro Systems. Front. Mar. Sci.6, 462 (2019). [Google Scholar]
11.Molari, M., Janssen, F., Vonnahme, T. R., Wenzhöfer, F. & Boetius, A. The contribution of microbial communities in polymetallic nodules to the diversity of the deep-sea microbiome of the Peru Basin (4130–4198 m depth). Biogeosciences17, 3203–3222 (2020). [Google Scholar]
12.Blöthe, M. et al. Manganese-Cycling Microbial Communities Inside Deep-Sea Manganese Nodules. Environ. Sci. Technol.49, 7692–7700 (2015). [DOI] [PubMed] [Google Scholar]
13.Danovaro, R. et al. Major viral impact on the functioning of benthic deep-sea ecosystems. Nature454, 1084–1087 (2008). [DOI] [PubMed] [Google Scholar]
14.Suttle, C. A. Marine viruses — major players in the global ecosystem. Nat. Rev. Microbiol.5, 801–812 (2007). [DOI] [PubMed] [Google Scholar]
15.Breitbart, M. Phage puppet masters of the marine microbial realm. Nat. Microbiol.3, 754–766 (2018). [DOI] [PubMed] [Google Scholar]
16.Cheng, R. et al. Virus diversity and interactions with hosts in deep-sea hydrothermal vents. Microbiome10, 1–17 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Li, Z. et al. Deep sea sediments associated with cold seeps are a subsurface reservoir of viral diversity. ISME J.15, 2366–2378 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Yu, M. et al. Diversity and potential host-interactions of viruses inhabiting deep-sea seamount sediments. Nat. Commun.15, 3228 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Zhao, J. et al. Novel Viral Communities Potentially Assisting in Carbon, Nitrogen, and Sulfur Metabolism in the Upper Slope Sediments of Mariana Trench. mSystems7, e01358–21 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Wilhelm, S. W. & Suttle, C. A. Viruses and Nutrient Cycles in the Sea: Viruses play critical roles in the structure and function of aquatic food webs. BioScience49, 781–788 (1999). [Google Scholar]
21.Breitbart, M. Marine Viruses: Truth or Dare. Annu. Rev. Mar. Sci.4, 425–448 (2012). [DOI] [PubMed] [Google Scholar]
22.Luo, X.-Q. et al. Viral community-wide auxiliary metabolic genes differ by lifestyles, habitats, and hosts. Microbiome10, 190 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Tian, F. et al. Prokaryotic-virus-encoded auxiliary metabolic genes throughout the global oceans. Microbiome12, 159 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Chen, Y. et al. Diversity and ecological potentials of viral assemblages from the seamount sediments of the Northwest Pacific Ocean. Innov. Geosci.2, 11 (2024). [Google Scholar]
25.Huang, D. et al. Enhanced mutualistic symbiosis between soil phages and bacteria with elevated chromium-induced environmental stress. Microbiome9, 150 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Chen, T. et al. Metagenomic analysis unveils the underexplored roles of prokaryotic viruses in a full-scale landfill leachate treatment plant. Water Res245, 120611 (2023). [DOI] [PubMed] [Google Scholar]
27.Wu, Z. et al. Distinct adaptive strategies and microbial interactions of soil viruses under different metal(loid) contaminations. J. Haz. Mat.460, 132347 (2023). [DOI] [PubMed] [Google Scholar]
28.Roux, S. et al. Minimum Information about an Uncultivated Virus Genome (MIUViG). Nat. Biotechnol.37, 29–37 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Al-Shayeb, B. et al. Clades of huge phages from across Earth’s ecosystems. Nature578, 425–431 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Gregory, A. C. et al. Marine DNA Viral Macro- and Microdiversity from Pole to Pole. Cell177, 1109–1123.e14 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Wu, Z. et al. Unveiling the unknown viral world in groundwater. Nat. Commun.15, 6788 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Nayfach, S. et al. Metagenomic compendium of 189,680 DNA viruses from the human gut microbiome. Nat. Microbiol.6, 960–970 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Low, S. J., Džunková, M., Chaumeil, P.-A., Parks, D. H. & Hugenholtz, P. Evaluation of a concatenated protein phylogeny for classification of tailed double-stranded DNA viruses belonging to the order Caudovirales. Nat. Microbiol.4, 1306–1315 (2019). [DOI] [PubMed] [Google Scholar]
34.Harding, K. R. et al. Genome-wide identification of bacterial genes contributing to nucleus-forming jumbo phage infection. Nucleic Acids Res53, gkae1194 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Jian, H. et al. Diversity and distribution of viruses inhabiting the deepest ocean on Earth. ISME J.15, 3094–3110 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Langwig, M. V. et al. Endemism shapes viral ecology and evolution in globally distributed hydrothermal vent ecosystems. Nat. Commun.16, 4076 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Jang, H. et al. Taxonomic assignment of uncultivated prokaryotic virus genomes is enabled by gene-sharing networks. Nat. Biotechnol.37, 632–639 (2019). [DOI] [PubMed] [Google Scholar]
38.Emerson, J. B. et al. Host-linked soil viral ecology along a permafrost thaw gradient. Nat. Microbiol.3, 870–880 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Payet, J. P. & Suttle, C. A. To kill or not to kill: The balance between lytic and lysogenic viral infection is driven by trophic status. Limnol. Oceanogr.58, 465–474 (2013). [Google Scholar]
40.Liu, S. et al. Close interactions between prokaryotes and plasmids or viruses highlight a pivotal role of horizontal gene transfer in shaping antibiotic/metal(loid) resistome and their prokaryotic supercarriers in untreated hospital sewage. Water Res286, 124178 (2025). [DOI] [PubMed] [Google Scholar]
41.Paez-Espino, D. et al. Uncovering Earth’s virome. Nature536, 425–430 (2016). [DOI] [PubMed] [Google Scholar]
42.Lever, M. A. et al. Life under extreme energy limitation: a synthesis of laboratory- and field-based investigations. FEMS Microbiol. Rev.39, 688–728 (2015). [DOI] [PubMed] [Google Scholar]
43.Gozzelino, R., Jeney, V. & Soares, M. P. Mechanisms of Cell Protection by Heme Oxygenase-1. Annu. Rev. Pharmacol. Toxicol.50, 323–354 (2010). [DOI] [PubMed] [Google Scholar]
44.Dammeyer, T., Bagby, S. C., Sullivan, M. B., Chisholm, S. W. & Frankenberg-Dinkel, N. Efficient Phage-Mediated Pigment Biosynthesis in Oceanic Cyanobacteria. Current Biology18, 442–448 (2008). [DOI] [PubMed]
45.Keffer, J. L. et al. Iron Oxidation by a Fused Cytochrome-Porin Common to Diverse Iron-Oxidizing Bacteria. mBio12, e01074-21 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Thompson, L. R. et al. Phage auxiliary metabolic genes and the redirection of cyanobacterial host carbon metabolism. Proc. Natl. Acad. Sci. USA108, E757–E764 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Nunoura, T. et al. Hadal biosphere: Insight into the microbial ecosystem in the deepest ocean on Earth. Proc. Natl. Acad. Sci. USA112, E1230–E1236 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Ruff, S. E. et al. Global dispersion and local diversification of the methane seep microbiome. Proc. Natl. Acad. Sci. USA112, 4015–4020 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Zhou, Z., St. John, E., Anantharaman, K. & Reysenbach, A.-L. Global patterns of diversity and metabolism of microbial communities in deep-sea hydrothermal vent deposits. Microbiome10, 241 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Mogollón, J. M., Mewes, K. & Kasten, S. Quantifying manganese and nitrogen cycle coupling in manganese-rich, organic carbon-starved marine sediments: Examples from the Clarion-Clipperton fracture zone. Geophys. Res. Lett.43, 7114–7123 (2016). [Google Scholar]
51.Guo, N. et al. Environmental drivers and ecological processes underlying microbial community zonation in deep-sea polymetallic nodule sediments. Mar. Environ. Res.215, 107790 (2026). [DOI] [PubMed] [Google Scholar]
52.Macomber, L. & Hausinger, R. P. Mechanisms of nickel toxicity in microorganisms. Metallomics3, 1153–1162 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Chandrangsu, P., Rensing, C. & Helmann, J. D. Metal homeostasis and resistance in bacteria. Nat. Rev. Microbiol.15, 338–350 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Zhang, H. et al. Dissecting the metal resistance genes contributed by virome from mining-affected metal contaminated soils. Front. Environ. Sci.11, 1182673 (2023). [Google Scholar]
55.Roe, K. L., Hogle, S. L. & Barbeau, K. A. Utilization of Heme as an Iron Source by Marine Alphaproteobacteria in the Roseobacter Clade. Appl. Environ. Microbiol.79, 18 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Huang, W. & Wilks, A. Extracellular Heme Uptake and the Challenge of Bacterial Cell Membranes. Annu. Rev. Biochem.86, 799–823 (2017). [DOI] [PubMed] [Google Scholar]
57.Cornell, R. M. & Schwertmann, U. The Iron Oxides: Structure, Properties, Reactions, Occurrences and Uses. 2nd ed. (Weinheim: Wiley-VCH, 2003).
58.Sweetman, A. K. et al. Key role of bacteria in the short-term cycling of carbon at the abyssal seafloor in a low particulate organic carbon flux region of the eastern Pacific Ocean. Limnol. Oceanogr.64, 694–713 (2019). [Google Scholar]
59.Smith, K. L. et al. Climate, carbon cycling, and deep-ocean ecosystems. Proc. Natl. Acad. Sci. USA106, 19211–19218 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Thomason, L. C., Court, D. L., Datta, A. R., Khanna, R. & Rosner, J. L. Identification of the Escherichia coli K-12 ybhE gene as pgl, encoding 6-phosphogluconolactonase. J. Bacteriol.186, 24 (2004). [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Hoehler, T. M. & Jørgensen, B. B. Microbial life under extreme energy limitation. Nat. Rev. Microbiol.11, 83–94 (2013). [DOI] [PubMed] [Google Scholar]
62.Hein, J. R., Koschinsky, A. & Kuhn, T. Deep-ocean polymetallic nodules as a resource for critical materials. Nat. Rev. Earth. Environ.1, 158–169 (2020). [Google Scholar]
63.Gazis, I.-Z. et al. Monitoring benthic plumes, sediment redeposition and seafloor imprints caused by deep-sea polymetallic nodule mining. Nat. Commun.16, 1229 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Vonnahme, T. R. et al. Effects of a deep-sea mining experiment on seafloor microbial communities and functions after 26 years. Sci. Adv.6, eaaz5922 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Magoč, T. & Salzberg, S. L. FLASH: fast length adjustment of short reads to improve genome assemblies. Bioinformatics27, 2957–2963 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Bolyen, E. et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat. Biotechnol.37, 852–857 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Chen, S., Zhou, Y., Chen, Y. & Gu, J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics34, i884–i890 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
68.Li, D., Liu, C.-M., Luo, R., Sadakane, K. & Lam, T.-W. MEGAHIT: an ultra-fast single-node solution for large and complex metagenomics assembly via succinct de Bruijn graph. Bioinformatics31, 1674–1676 (2015). [DOI] [PubMed] [Google Scholar]
69.Uritskiy, G. V., DiRuggiero, J. & Taylor, J. MetaWRAP—a flexible pipeline for genome-resolved metagenomic data analysis. Microbiome6, 158 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
70.Parks, D. H., Imelfort, M., Skennerton, C. T., Hugenholtz, P. & Tyson, G. W. CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes. Genome Res25, 1043–1055 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
71.The Genome Standards Consortium 2017.Minimum information about a single amplified genome (MISAG) and a metagenome-assembled genome (MIMAG) of bacteria and archaea Nat. Biotechnol.35, 725–731 (2017). [DOI] [PMC free article] [PubMed]
72.Olm, M. R., Brown, C. T., Brooks, B. & Banfield, J. F. dRep: a tool for fast and accurate genomic comparisons that enables improved genome recovery from metagenomes through de-replication. ISME J.11, 2864–2868 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
73.Chaumeil, P.-A., Mussig, A. J., Hugenholtz, P. & Parks, D. H. GTDB-Tk v2: memory-friendly classification with the genome taxonomy database. Bioinformatics38, 5315–5316 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
74.Price, M. N., Dehal, P. S. & Arkin, A. P. FastTree 2 – approximately maximum-likelihood trees for large alignments. PLoS ONE5, e9490 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
75.Letunic, I. & Bork, P. Interactive Tree of Life (iTOL) v6: recent updates to the phylogenetic tree display and annotation tool. Nucleic Acids Res52, W78–W82 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
76.Zhou, Z. et al. METABOLIC: high-throughput profiling of microbial genomes for functional traits, metabolism, biogeochemistry, and community-scale functional networks. Microbiome10, 33 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
77.Paez-Espino, D., Pavlopoulos, G. A., Ivanova, N. N. & Kyrpides, N. C. Nontargeted virus sequence discovery pipeline and virus clustering for metagenomic data. Nat. Protoc.12, 1673–1682 (2017). [DOI] [PubMed] [Google Scholar]
78.Kieft, K., Zhou, Z. & Anantharaman, K. VIBRANT: automated recovery, annotation and curation of microbial viruses, and evaluation of viral community function from genomic sequences. Microbiome8, 90 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
79.Guo, J. et al. VirSorter2: a multi-classifier, expert-guided approach to detect diverse DNA and RNA viruses. Microbiome9, 37 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
80.Ren, J. et al. Identifying viruses from metagenomic data using deep learning. Quant. Biol.8, 64–77 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
81.Fang, Z. et al. PPR-Meta: a tool for identifying phages and plasmids from metagenomic fragments using deep learning. GigaScience8, giz066 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
82.Nayfach, S. et al. CheckV assesses the quality and completeness of metagenome-assembled viral genomes. Nat. Biotechnol.39, 578–585 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
83.Ye, J., McGinnis, S. & Madden, T. L. BLAST: improvements for better sequence analysis. Nucleic Acids Res.34, W6–W9 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]
84.Finn, R. D., Clements, J. & Eddy, S. R. HMMER web server: interactive sequence similarity searching. Nucleic Acids Res.39, W29–W37 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
85.Cook, R. et al. Hybrid assembly of an agricultural slurry virome reveals a diverse and stable community with the potential to alter the metabolism and virulence of veterinary pathogens. Microbiome9, 65 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
86.Hyatt, D. et al. Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinforma.11, 119 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
87.Fu, L., Niu, B., Zhu, Z., Wu, S. & Li, W. CD-HIT: accelerated for clustering the next-generation sequencing data. Bioinformatics28, 3150–3152 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
88.Camargo, A. P. et al. Identification of mobile genetic elements with geNomad. Nat. Biotechnol.42, 1303–1312 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
89.Li, H. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. Preprint at 10.48550/arXiv.1303.3997 (2013). [DOI]
90.Li, H. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics25, 2078–2079 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
91.Mende, D. R. et al. Environmental drivers of a microbial genomic transition zone in the ocean’s interior. Nat. Microbiol.2, 1367–1373 (2017). [DOI] [PubMed] [Google Scholar]
92.Roux, S. et al. iPHoP: An integrated machine learning framework to maximize host prediction for metagenome-derived viruses of archaea and bacteria. PLoS Biol.21, e3002083 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
93.Shaffer, M. et al. DRAM for distilling microbial metabolism to automate the curation of microbiome function. Nucleic Acids Res48, 8883–8900 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
94.Martin, C., Emerson, J. B., Roux, S. & Anantharaman, K. A call for caution in the biological interpretation of viral auxiliary metabolic genes. Nat. Microbiol.10, 2122–2129 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
95.Mirdita, M. et al. ColabFold: making protein folding accessible to all. Nat. Methods19, 679–682 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
96.Pettersen, E. F. et al. UCSF CHIMERAX: Structure visualization for researchers, educators, and developers. Protein Sci.30, 70–82 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
97.Kelley, L. A., Mezulis, S., Yates, C. M., Wass, M. N. & Sternberg, M. J. E. The Phyre2 web portal for protein modeling, prediction and analysis. Nat. Protoc.10, 845–858 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
98.Buchfink, B., Reuter, K. & Drost, H.-G. Sensitive protein alignments at tree-of-life scale using DIAMOND. Nat. Methods18, 366–368 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
99.Shannon, P. et al. Cytoscape: A Software Environment for Integrated Models of Biomolecular Interaction Networks. Genome Res13, 2498–2504 (2003). [DOI] [PMC free article] [PubMed] [Google Scholar]
100.Gregory, A. C. et al. MetaPop: a pipeline for macro- and microdiversity analyses and visualization of microbial and viral metagenome-derived populations. Microbiome10, 49 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
101.Capella-Gutiérrez, S., Silla-Martínez, J. M. & Gabaldón, T. trimAl: a tool for automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics25, 1972–1973 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
102.Edgar, R. C. MUSCLE: a multiple sequence alignment method with reduced time and space complexity. BMC Bioinforma.5, 113 (2004). [DOI] [PMC free article] [PubMed] [Google Scholar]
103.Zhou, Y.-L., Mara, P., Cui, G.-J., Edgcomb, V. P. & Wang, Y. Microbiomes in the Challenger Deep slope and bottom-axis sediments. Nat. Commun.13, 1515 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
104.Li, Y., Liu, H., Xiao, Y. & Jing, H. Metagenome sequencing and 982 microbial genomes from Kermadec and Diamantina Trenches sediments. Sci. Data11, 1067 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
105.Pal, C., Bengtsson-Palme, J., Rensing, C., Kristiansson, E. & Larsson, D. G. J. BacMet: antibacterial biocide and metal resistance genes database. Nucleic Acids Res.42, D737–D743 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
106.Ogata, H. et al. Genome sequence of Rickettsia bellii illuminates the role of amoebae in gene exchanges between intracellular pathogens. PLoS Genet.2, e76 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]
107.R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, Vienna, Austria, 2024).
108.Dixon, P. VEGAN, a package of R functions for community ecology. Veg. Sci.14, 927–930 (2003). [Google Scholar]
109.Kembel, S. W. et al. Picante: R tools for integrating phylogenies and ecology. Bioinformatics26, 1463–1464 (2010). [DOI] [PubMed] [Google Scholar]
110.Bates, D., Mächler, M., Bolker, B. & Walker, S. Fitting linear mixed-effects models using lme4. Stat. Softw.67, 1–48 (2015). [Google Scholar]
111.Greenacre, M. Compositional data analysis. Annu. Rev. Stat. Appl.8, 271–299 (2021). [Google Scholar]
112.Quinn, T. P. et al. A field guide for the compositional analysis of any-omics data. GigaScience8, giz107 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
113.Kosmopoulos, J. C., Pallier, W., Malik, A. A. & Anantharaman, K. Ecosystem health shapes viral ecology in peatland soils. Nat. Microbiol.11, 142–154 (2026). [DOI] [PMC free article] [PubMed] [Google Scholar]
114.Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol.15, 550 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
115.Gómez-Rubio, V. ggplot2 - Elegant graphics for data analysis (2nd Edition). Stat. Softw.77, 1–3 (2017). [Google Scholar]
116.Chen, H. & Boutros, P. C. VennDiagram: a package for the generation of highly-customizable Venn and Euler diagrams in R. BMC Bioinforma.12, 35 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
117.Conway, J. R., Lex, A. & Gehlenborg, N. UpSetR: an R package for the visualization of intersecting sets and their properties. Bioinformatics33, 2938–2940 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
118.QGIS Development Team. Geographic Information System v. 3.40.7 (Open-Source Geospatial Foundation Project, 2025).
119.GEBCO Compilation Group. GEBCO 2025 Grid 10.5285/37c52e96-24ea-67ce-e063-7086abc05f29 (2025). [DOI]
120.Gios, E., Mosley, O. E., Hoggard, M. & Handley, K. M. High niche specificity and host genetic diversity of groundwater viruses. ISME J.18, wrae035 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information^{(29.8MB, pdf)}

41467_2026_78161_MOESM2_ESM.pdf^{(6.1KB, pdf)}

Description of Additional Supplementary Files

Supplementary Data 1-11^{(16.4MB, xlsx)}

Reporting Summary^{(3.3MB, pdf)}

Transparent Peer Review file^{(35.5MB, pdf)}

Source Data^{(4.4MB, xlsx)}

Data Availability Statement

The custom code used for viral identification and figure generation are available at https://github.com/Bowencqu/CCZV/ or https://doi.org/10.5281/zenodo.21306299.

[CR1] 1.Chen, J. et al. Global marine microbial diversity and its potential in bioprospecting. Nature633, 371–379 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR2] 2.Zhang, C., Peng, Y., Liu, X., Wang, J. & Dong, X. Deep-sea microbial genetic resources: new frontiers for bioprospecting. Trends Microbiol32, 321–324 (2024). [DOI] [PubMed] [Google Scholar]

[CR3] 3.Hein, J. R., Mizell, K., Koschinsky, A. & Conrad, T. A. Deep-ocean mineral deposits as a source of critical metals for high- and green-technology applications: comparison with land-based resources. Ore Geol. Rev.51, 1–14 (2013). [Google Scholar]

[CR4] 4.Koschinsky, A. & Hein, J. R. Marine Ferromanganese Encrustations: Archives of Changing Oceans. Elements13, 177–182 (2017). [Google Scholar]

[CR5] 5.Simon-Lledó, E. et al. Ecology of a polymetallic nodule occurrence gradient: Implications for deep-sea mining. Limnol. Oceanogr.64, 1883–1894 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR6] 6.Shiraishi, F. et al. Dense microbial community on a ferromanganese nodule from the ultra-oligotrophic South Pacific Gyre: Implications for biogeochemical cycles. Earth Planet. Sci. Lett.447, 10–20 (2016). [Google Scholar]

[CR7] 7.Versteegh, G. J. M., Koschinsky, A., Kuhn, T., Preuss, I. & Kasten, S. Geochemical consequences of oxygen diffusion from the oceanic crust into overlying sediments and its significance for biogeochemical cycles based on sediments of the northeast Pacific. Biogeosciences18, 4965–4984 (2021). [Google Scholar]

[CR8] 8.China Minmetals Corporation. Environmental Impact Statement (China Minmetals Corporation, 2025).

[CR9] 9.Zhang, D. et al. Microbe-driven elemental cycling enables microbial adaptation to deep-sea ferromanganese nodule sediment fields. Microbiome11, 160 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR10] 10.Gillard, B., Chatzievangelou, D., Thomsen, L. & Ullrich, M. S. Heavy-Metal-Resistant Microorganisms in Deep-Sea Sediments Disturbed by Mining Activity: An Application Toward the Development of Experimental in vitro Systems. Front. Mar. Sci.6, 462 (2019). [Google Scholar]

[CR11] 11.Molari, M., Janssen, F., Vonnahme, T. R., Wenzhöfer, F. & Boetius, A. The contribution of microbial communities in polymetallic nodules to the diversity of the deep-sea microbiome of the Peru Basin (4130–4198 m depth). Biogeosciences17, 3203–3222 (2020). [Google Scholar]

[CR12] 12.Blöthe, M. et al. Manganese-Cycling Microbial Communities Inside Deep-Sea Manganese Nodules. Environ. Sci. Technol.49, 7692–7700 (2015). [DOI] [PubMed] [Google Scholar]

[CR13] 13.Danovaro, R. et al. Major viral impact on the functioning of benthic deep-sea ecosystems. Nature454, 1084–1087 (2008). [DOI] [PubMed] [Google Scholar]

[CR14] 14.Suttle, C. A. Marine viruses — major players in the global ecosystem. Nat. Rev. Microbiol.5, 801–812 (2007). [DOI] [PubMed] [Google Scholar]

[CR15] 15.Breitbart, M. Phage puppet masters of the marine microbial realm. Nat. Microbiol.3, 754–766 (2018). [DOI] [PubMed] [Google Scholar]

[CR16] 16.Cheng, R. et al. Virus diversity and interactions with hosts in deep-sea hydrothermal vents. Microbiome10, 1–17 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR17] 17.Li, Z. et al. Deep sea sediments associated with cold seeps are a subsurface reservoir of viral diversity. ISME J.15, 2366–2378 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR18] 18.Yu, M. et al. Diversity and potential host-interactions of viruses inhabiting deep-sea seamount sediments. Nat. Commun.15, 3228 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR19] 19.Zhao, J. et al. Novel Viral Communities Potentially Assisting in Carbon, Nitrogen, and Sulfur Metabolism in the Upper Slope Sediments of Mariana Trench. mSystems7, e01358–21 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR20] 20.Wilhelm, S. W. & Suttle, C. A. Viruses and Nutrient Cycles in the Sea: Viruses play critical roles in the structure and function of aquatic food webs. BioScience49, 781–788 (1999). [Google Scholar]

[CR21] 21.Breitbart, M. Marine Viruses: Truth or Dare. Annu. Rev. Mar. Sci.4, 425–448 (2012). [DOI] [PubMed] [Google Scholar]

[CR22] 22.Luo, X.-Q. et al. Viral community-wide auxiliary metabolic genes differ by lifestyles, habitats, and hosts. Microbiome10, 190 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR23] 23.Tian, F. et al. Prokaryotic-virus-encoded auxiliary metabolic genes throughout the global oceans. Microbiome12, 159 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR24] 24.Chen, Y. et al. Diversity and ecological potentials of viral assemblages from the seamount sediments of the Northwest Pacific Ocean. Innov. Geosci.2, 11 (2024). [Google Scholar]

[CR25] 25.Huang, D. et al. Enhanced mutualistic symbiosis between soil phages and bacteria with elevated chromium-induced environmental stress. Microbiome9, 150 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR26] 26.Chen, T. et al. Metagenomic analysis unveils the underexplored roles of prokaryotic viruses in a full-scale landfill leachate treatment plant. Water Res245, 120611 (2023). [DOI] [PubMed] [Google Scholar]

[CR27] 27.Wu, Z. et al. Distinct adaptive strategies and microbial interactions of soil viruses under different metal(loid) contaminations. J. Haz. Mat.460, 132347 (2023). [DOI] [PubMed] [Google Scholar]

[CR28] 28.Roux, S. et al. Minimum Information about an Uncultivated Virus Genome (MIUViG). Nat. Biotechnol.37, 29–37 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR29] 29.Al-Shayeb, B. et al. Clades of huge phages from across Earth’s ecosystems. Nature578, 425–431 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR30] 30.Gregory, A. C. et al. Marine DNA Viral Macro- and Microdiversity from Pole to Pole. Cell177, 1109–1123.e14 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR31] 31.Wu, Z. et al. Unveiling the unknown viral world in groundwater. Nat. Commun.15, 6788 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR32] 32.Nayfach, S. et al. Metagenomic compendium of 189,680 DNA viruses from the human gut microbiome. Nat. Microbiol.6, 960–970 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR33] 33.Low, S. J., Džunková, M., Chaumeil, P.-A., Parks, D. H. & Hugenholtz, P. Evaluation of a concatenated protein phylogeny for classification of tailed double-stranded DNA viruses belonging to the order Caudovirales. Nat. Microbiol.4, 1306–1315 (2019). [DOI] [PubMed] [Google Scholar]

[CR34] 34.Harding, K. R. et al. Genome-wide identification of bacterial genes contributing to nucleus-forming jumbo phage infection. Nucleic Acids Res53, gkae1194 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR35] 35.Jian, H. et al. Diversity and distribution of viruses inhabiting the deepest ocean on Earth. ISME J.15, 3094–3110 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR36] 36.Langwig, M. V. et al. Endemism shapes viral ecology and evolution in globally distributed hydrothermal vent ecosystems. Nat. Commun.16, 4076 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR37] 37.Jang, H. et al. Taxonomic assignment of uncultivated prokaryotic virus genomes is enabled by gene-sharing networks. Nat. Biotechnol.37, 632–639 (2019). [DOI] [PubMed] [Google Scholar]

[CR38] 38.Emerson, J. B. et al. Host-linked soil viral ecology along a permafrost thaw gradient. Nat. Microbiol.3, 870–880 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR39] 39.Payet, J. P. & Suttle, C. A. To kill or not to kill: The balance between lytic and lysogenic viral infection is driven by trophic status. Limnol. Oceanogr.58, 465–474 (2013). [Google Scholar]

[CR40] 40.Liu, S. et al. Close interactions between prokaryotes and plasmids or viruses highlight a pivotal role of horizontal gene transfer in shaping antibiotic/metal(loid) resistome and their prokaryotic supercarriers in untreated hospital sewage. Water Res286, 124178 (2025). [DOI] [PubMed] [Google Scholar]

[CR41] 41.Paez-Espino, D. et al. Uncovering Earth’s virome. Nature536, 425–430 (2016). [DOI] [PubMed] [Google Scholar]

[CR42] 42.Lever, M. A. et al. Life under extreme energy limitation: a synthesis of laboratory- and field-based investigations. FEMS Microbiol. Rev.39, 688–728 (2015). [DOI] [PubMed] [Google Scholar]

[CR43] 43.Gozzelino, R., Jeney, V. & Soares, M. P. Mechanisms of Cell Protection by Heme Oxygenase-1. Annu. Rev. Pharmacol. Toxicol.50, 323–354 (2010). [DOI] [PubMed] [Google Scholar]

[CR44] 44.Dammeyer, T., Bagby, S. C., Sullivan, M. B., Chisholm, S. W. & Frankenberg-Dinkel, N. Efficient Phage-Mediated Pigment Biosynthesis in Oceanic Cyanobacteria. Current Biology18, 442–448 (2008). [DOI] [PubMed]

[CR45] 45.Keffer, J. L. et al. Iron Oxidation by a Fused Cytochrome-Porin Common to Diverse Iron-Oxidizing Bacteria. mBio12, e01074-21 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR46] 46.Thompson, L. R. et al. Phage auxiliary metabolic genes and the redirection of cyanobacterial host carbon metabolism. Proc. Natl. Acad. Sci. USA108, E757–E764 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR47] 47.Nunoura, T. et al. Hadal biosphere: Insight into the microbial ecosystem in the deepest ocean on Earth. Proc. Natl. Acad. Sci. USA112, E1230–E1236 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR48] 48.Ruff, S. E. et al. Global dispersion and local diversification of the methane seep microbiome. Proc. Natl. Acad. Sci. USA112, 4015–4020 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR49] 49.Zhou, Z., St. John, E., Anantharaman, K. & Reysenbach, A.-L. Global patterns of diversity and metabolism of microbial communities in deep-sea hydrothermal vent deposits. Microbiome10, 241 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR50] 50.Mogollón, J. M., Mewes, K. & Kasten, S. Quantifying manganese and nitrogen cycle coupling in manganese-rich, organic carbon-starved marine sediments: Examples from the Clarion-Clipperton fracture zone. Geophys. Res. Lett.43, 7114–7123 (2016). [Google Scholar]

[CR51] 51.Guo, N. et al. Environmental drivers and ecological processes underlying microbial community zonation in deep-sea polymetallic nodule sediments. Mar. Environ. Res.215, 107790 (2026). [DOI] [PubMed] [Google Scholar]

[CR52] 52.Macomber, L. & Hausinger, R. P. Mechanisms of nickel toxicity in microorganisms. Metallomics3, 1153–1162 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR53] 53.Chandrangsu, P., Rensing, C. & Helmann, J. D. Metal homeostasis and resistance in bacteria. Nat. Rev. Microbiol.15, 338–350 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR54] 54.Zhang, H. et al. Dissecting the metal resistance genes contributed by virome from mining-affected metal contaminated soils. Front. Environ. Sci.11, 1182673 (2023). [Google Scholar]

[CR55] 55.Roe, K. L., Hogle, S. L. & Barbeau, K. A. Utilization of Heme as an Iron Source by Marine Alphaproteobacteria in the Roseobacter Clade. Appl. Environ. Microbiol.79, 18 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR56] 56.Huang, W. & Wilks, A. Extracellular Heme Uptake and the Challenge of Bacterial Cell Membranes. Annu. Rev. Biochem.86, 799–823 (2017). [DOI] [PubMed] [Google Scholar]

[CR57] 57.Cornell, R. M. & Schwertmann, U. The Iron Oxides: Structure, Properties, Reactions, Occurrences and Uses. 2nd ed. (Weinheim: Wiley-VCH, 2003).

[CR58] 58.Sweetman, A. K. et al. Key role of bacteria in the short-term cycling of carbon at the abyssal seafloor in a low particulate organic carbon flux region of the eastern Pacific Ocean. Limnol. Oceanogr.64, 694–713 (2019). [Google Scholar]

[CR59] 59.Smith, K. L. et al. Climate, carbon cycling, and deep-ocean ecosystems. Proc. Natl. Acad. Sci. USA106, 19211–19218 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR60] 60.Thomason, L. C., Court, D. L., Datta, A. R., Khanna, R. & Rosner, J. L. Identification of the Escherichia coli K-12 ybhE gene as pgl, encoding 6-phosphogluconolactonase. J. Bacteriol.186, 24 (2004). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR61] 61.Hoehler, T. M. & Jørgensen, B. B. Microbial life under extreme energy limitation. Nat. Rev. Microbiol.11, 83–94 (2013). [DOI] [PubMed] [Google Scholar]

[CR62] 62.Hein, J. R., Koschinsky, A. & Kuhn, T. Deep-ocean polymetallic nodules as a resource for critical materials. Nat. Rev. Earth. Environ.1, 158–169 (2020). [Google Scholar]

[CR63] 63.Gazis, I.-Z. et al. Monitoring benthic plumes, sediment redeposition and seafloor imprints caused by deep-sea polymetallic nodule mining. Nat. Commun.16, 1229 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR64] 64.Vonnahme, T. R. et al. Effects of a deep-sea mining experiment on seafloor microbial communities and functions after 26 years. Sci. Adv.6, eaaz5922 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR65] 65.Magoč, T. & Salzberg, S. L. FLASH: fast length adjustment of short reads to improve genome assemblies. Bioinformatics27, 2957–2963 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR66] 66.Bolyen, E. et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat. Biotechnol.37, 852–857 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR67] 67.Chen, S., Zhou, Y., Chen, Y. & Gu, J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics34, i884–i890 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR68] 68.Li, D., Liu, C.-M., Luo, R., Sadakane, K. & Lam, T.-W. MEGAHIT: an ultra-fast single-node solution for large and complex metagenomics assembly via succinct de Bruijn graph. Bioinformatics31, 1674–1676 (2015). [DOI] [PubMed] [Google Scholar]

[CR69] 69.Uritskiy, G. V., DiRuggiero, J. & Taylor, J. MetaWRAP—a flexible pipeline for genome-resolved metagenomic data analysis. Microbiome6, 158 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR70] 70.Parks, D. H., Imelfort, M., Skennerton, C. T., Hugenholtz, P. & Tyson, G. W. CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes. Genome Res25, 1043–1055 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR71] 71.The Genome Standards Consortium 2017.Minimum information about a single amplified genome (MISAG) and a metagenome-assembled genome (MIMAG) of bacteria and archaea Nat. Biotechnol.35, 725–731 (2017). [DOI] [PMC free article] [PubMed]

[CR72] 72.Olm, M. R., Brown, C. T., Brooks, B. & Banfield, J. F. dRep: a tool for fast and accurate genomic comparisons that enables improved genome recovery from metagenomes through de-replication. ISME J.11, 2864–2868 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR73] 73.Chaumeil, P.-A., Mussig, A. J., Hugenholtz, P. & Parks, D. H. GTDB-Tk v2: memory-friendly classification with the genome taxonomy database. Bioinformatics38, 5315–5316 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR74] 74.Price, M. N., Dehal, P. S. & Arkin, A. P. FastTree 2 – approximately maximum-likelihood trees for large alignments. PLoS ONE5, e9490 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR75] 75.Letunic, I. & Bork, P. Interactive Tree of Life (iTOL) v6: recent updates to the phylogenetic tree display and annotation tool. Nucleic Acids Res52, W78–W82 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR76] 76.Zhou, Z. et al. METABOLIC: high-throughput profiling of microbial genomes for functional traits, metabolism, biogeochemistry, and community-scale functional networks. Microbiome10, 33 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR77] 77.Paez-Espino, D., Pavlopoulos, G. A., Ivanova, N. N. & Kyrpides, N. C. Nontargeted virus sequence discovery pipeline and virus clustering for metagenomic data. Nat. Protoc.12, 1673–1682 (2017). [DOI] [PubMed] [Google Scholar]

[CR78] 78.Kieft, K., Zhou, Z. & Anantharaman, K. VIBRANT: automated recovery, annotation and curation of microbial viruses, and evaluation of viral community function from genomic sequences. Microbiome8, 90 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR79] 79.Guo, J. et al. VirSorter2: a multi-classifier, expert-guided approach to detect diverse DNA and RNA viruses. Microbiome9, 37 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR80] 80.Ren, J. et al. Identifying viruses from metagenomic data using deep learning. Quant. Biol.8, 64–77 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR81] 81.Fang, Z. et al. PPR-Meta: a tool for identifying phages and plasmids from metagenomic fragments using deep learning. GigaScience8, giz066 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR82] 82.Nayfach, S. et al. CheckV assesses the quality and completeness of metagenome-assembled viral genomes. Nat. Biotechnol.39, 578–585 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR83] 83.Ye, J., McGinnis, S. & Madden, T. L. BLAST: improvements for better sequence analysis. Nucleic Acids Res.34, W6–W9 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR84] 84.Finn, R. D., Clements, J. & Eddy, S. R. HMMER web server: interactive sequence similarity searching. Nucleic Acids Res.39, W29–W37 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR85] 85.Cook, R. et al. Hybrid assembly of an agricultural slurry virome reveals a diverse and stable community with the potential to alter the metabolism and virulence of veterinary pathogens. Microbiome9, 65 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR86] 86.Hyatt, D. et al. Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinforma.11, 119 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR87] 87.Fu, L., Niu, B., Zhu, Z., Wu, S. & Li, W. CD-HIT: accelerated for clustering the next-generation sequencing data. Bioinformatics28, 3150–3152 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR88] 88.Camargo, A. P. et al. Identification of mobile genetic elements with geNomad. Nat. Biotechnol.42, 1303–1312 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR89] 89.Li, H. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. Preprint at 10.48550/arXiv.1303.3997 (2013). [DOI]

[CR90] 90.Li, H. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics25, 2078–2079 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR91] 91.Mende, D. R. et al. Environmental drivers of a microbial genomic transition zone in the ocean’s interior. Nat. Microbiol.2, 1367–1373 (2017). [DOI] [PubMed] [Google Scholar]

[CR92] 92.Roux, S. et al. iPHoP: An integrated machine learning framework to maximize host prediction for metagenome-derived viruses of archaea and bacteria. PLoS Biol.21, e3002083 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR93] 93.Shaffer, M. et al. DRAM for distilling microbial metabolism to automate the curation of microbiome function. Nucleic Acids Res48, 8883–8900 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR94] 94.Martin, C., Emerson, J. B., Roux, S. & Anantharaman, K. A call for caution in the biological interpretation of viral auxiliary metabolic genes. Nat. Microbiol.10, 2122–2129 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR95] 95.Mirdita, M. et al. ColabFold: making protein folding accessible to all. Nat. Methods19, 679–682 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR96] 96.Pettersen, E. F. et al. UCSF CHIMERAX: Structure visualization for researchers, educators, and developers. Protein Sci.30, 70–82 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR97] 97.Kelley, L. A., Mezulis, S., Yates, C. M., Wass, M. N. & Sternberg, M. J. E. The Phyre2 web portal for protein modeling, prediction and analysis. Nat. Protoc.10, 845–858 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR98] 98.Buchfink, B., Reuter, K. & Drost, H.-G. Sensitive protein alignments at tree-of-life scale using DIAMOND. Nat. Methods18, 366–368 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR99] 99.Shannon, P. et al. Cytoscape: A Software Environment for Integrated Models of Biomolecular Interaction Networks. Genome Res13, 2498–2504 (2003). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR100] 100.Gregory, A. C. et al. MetaPop: a pipeline for macro- and microdiversity analyses and visualization of microbial and viral metagenome-derived populations. Microbiome10, 49 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR101] 101.Capella-Gutiérrez, S., Silla-Martínez, J. M. & Gabaldón, T. trimAl: a tool for automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics25, 1972–1973 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR102] 102.Edgar, R. C. MUSCLE: a multiple sequence alignment method with reduced time and space complexity. BMC Bioinforma.5, 113 (2004). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR103] 103.Zhou, Y.-L., Mara, P., Cui, G.-J., Edgcomb, V. P. & Wang, Y. Microbiomes in the Challenger Deep slope and bottom-axis sediments. Nat. Commun.13, 1515 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR104] 104.Li, Y., Liu, H., Xiao, Y. & Jing, H. Metagenome sequencing and 982 microbial genomes from Kermadec and Diamantina Trenches sediments. Sci. Data11, 1067 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR105] 105.Pal, C., Bengtsson-Palme, J., Rensing, C., Kristiansson, E. & Larsson, D. G. J. BacMet: antibacterial biocide and metal resistance genes database. Nucleic Acids Res.42, D737–D743 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR106] 106.Ogata, H. et al. Genome sequence of Rickettsia bellii illuminates the role of amoebae in gene exchanges between intracellular pathogens. PLoS Genet.2, e76 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR107] 107.R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, Vienna, Austria, 2024).

[CR108] 108.Dixon, P. VEGAN, a package of R functions for community ecology. Veg. Sci.14, 927–930 (2003). [Google Scholar]

[CR109] 109.Kembel, S. W. et al. Picante: R tools for integrating phylogenies and ecology. Bioinformatics26, 1463–1464 (2010). [DOI] [PubMed] [Google Scholar]

[CR110] 110.Bates, D., Mächler, M., Bolker, B. & Walker, S. Fitting linear mixed-effects models using lme4. Stat. Softw.67, 1–48 (2015). [Google Scholar]

[CR111] 111.Greenacre, M. Compositional data analysis. Annu. Rev. Stat. Appl.8, 271–299 (2021). [Google Scholar]

[CR112] 112.Quinn, T. P. et al. A field guide for the compositional analysis of any-omics data. GigaScience8, giz107 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR113] 113.Kosmopoulos, J. C., Pallier, W., Malik, A. A. & Anantharaman, K. Ecosystem health shapes viral ecology in peatland soils. Nat. Microbiol.11, 142–154 (2026). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR114] 114.Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol.15, 550 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR115] 115.Gómez-Rubio, V. ggplot2 - Elegant graphics for data analysis (2nd Edition). Stat. Softw.77, 1–3 (2017). [Google Scholar]

[CR116] 116.Chen, H. & Boutros, P. C. VennDiagram: a package for the generation of highly-customizable Venn and Euler diagrams in R. BMC Bioinforma.12, 35 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR117] 117.Conway, J. R., Lex, A. & Gehlenborg, N. UpSetR: an R package for the visualization of intersecting sets and their properties. Bioinformatics33, 2938–2940 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR118] 118.QGIS Development Team. Geographic Information System v. 3.40.7 (Open-Source Geospatial Foundation Project, 2025).

[CR119] 119.GEBCO Compilation Group. GEBCO 2025 Grid 10.5285/37c52e96-24ea-67ce-e063-7086abc05f29 (2025). [DOI]

[CR120] 120.Gios, E., Mosley, O. E., Hoggard, M. & Handley, K. M. High niche specificity and host genetic diversity of groundwater viruses. ISME J.18, wrae035 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Viral diversity and ecological functions in the world’s largest deep-sea mineral exploration region

Bowen Hou

Lilan Zhang

Dong Sun

Abstract

Introduction

Results

Fig. 1. Geographic distribution of collected sediment samples in the Clarion-Clipperton Fracture Zone (CCZ).

Establishment of the Clarion-Clipperton fracture zone virome (CCZV)

Fig. 2. Overview of the Clarion-Clipperton Fracture Zone Virome (CCZV).

Depth-related patterns in viral communities, populations, and protein clusters

Cross-ecosystem comparison reveals extensive heterogeneity in CCZV viral lineages and proteins

Fig. 3. Cross-ecosystem comparison of CCZV viral population and viral protein repertoires.

Virus–host linkages in CCZ sediments

Fig. 4. Virus-host linkages in the CCZV.

Virus-carried metal resistance genes

Potential roles of viral AMGs in biogenic element cycles

Fig. 5. Function and abundance profiles of virus-encoded auxiliary metabolic genes (AMGs).

Potential involvement of viruses in metal cycling

Fig. 6. Potential viral impacts on metal cycling.

Discussion

Methods

Sample collection and element analysis

DNA extraction, 16S rRNA gene sequencing and analysis

Metagenomic sequencing and assembly

Reconstruction of prokaryotic metagenome-assembled genomes

Identification of viral contigs

Lifestyle prediction

Clustering of viral proteins and taxonomic annotation of vOTUs

Abundance profiles

Gene characteristic calculations for GC content, protein molecular weight, and protein elemental composition

Prediction of virus–host linkages

Functional annotation and AMG identification

Comparisons of viral genomes to public databases

Microdiversity analyses of viral genomes

Phylogenetic analyses

Metal resistance gene identification

Gene exchange analysis

Statistical analyses and plotting

Reporting summary

Supplementary information

Source data

Acknowledgements

Author contributions

Peer review

Peer review information

Funding

Data availability

Code availability

Competing interests

Footnotes

Contributor Information

Supplementary information

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases