Antarctic ice shelves, vast floating ice masses extending from the continent, significantly impact global carbon cycles. Previously considered inactive, recent research has revealed the presence of unique microbial communities driving biogeochemical cycles within these ice shelves. The Ross Ice Shelf (RIS), the largest in Antarctica, covers a massive area and possesses unique physico-chemical properties, making it a crucial component of the Antarctic ice system and a significant contributor to the global climate system. These polar ecosystems, while harboring remarkable microbial diversity, are increasingly threatened by climate change. A recent study characterized the microbial community beneath the RIS, identifying key primary producers like *Nitrosopumilus* spp. (ammonium-oxidizing archaea) and *Thioglobus* spp. (sulfur-oxidizing bacteria). The study highlighted the importance of understanding the role of viruses in these ecosystems given their established impact on marine ecosystem functioning and biogeochemical processes, particularly in the vast and relatively unexplored mesopelagic and bathypelagic zones of the ocean. This research aims to address the knowledge gap regarding viral diversity, activity, and ecological roles within the RIS, considering the significance of understanding these interactions in the context of climate change impacts on these unique ecosystems.

The role of marine viruses in global biogeochemical cycles is increasingly recognized, with studies demonstrating their influence on the biological carbon pump through viral "shunt" and "shuttle" mechanisms. Some studies have even linked the predicted infection of ecologically important hosts by viruses to a significant portion of the variation in organic carbon export. Meso- and bathypelagic marine environments, the largest ecosystems on Earth, are particularly crucial for understanding viral influence on the global carbon cycle. However, our understanding of viruses in dark ocean environments, especially under polar ice shelves, remains limited due to the significant challenges in sampling these remote and extreme habitats. Previous research on the RIS microbial community has highlighted the importance of chemolithoautotrophs in primary production, but the role of viruses in regulating these communities remained largely uninvestigated. Existing models such as the 'Piggyback-the-Winner' (PtW) hypothesis posit that phage integration into host genomes (as prophages) is favored in high-productivity environments, while alternative strategies like 'Piggyback-the-Persistent' suggest dominance of temperate phages in oligotrophic environments. Understanding these dynamics is critical for predicting the response of these ecosystems to climate change.

This study leveraged previously generated data from the RIS Program, which included single-cell genomics (SAGs), metagenomics, metatranscriptomics, and metagenome-assembled genomes (MAGs) from seawater samples collected at three depths (30, 180, and 330 m) beneath the RIS. Because the original RIS program did not specifically focus on free-living viral communities, the analysis used a conservative approach to identify viral sequences. The researchers used a combination of Virsorter 2.0, CheckV, and PPR-meta programs to identify and validate 607 bona fide viral genomes. Viral classification utilized both Virsorter 2.0 (identifying hallmark genes) and Genomad (using the ICTV database). Biogeography was analyzed by comparing the RIS viral genes with those from the GOV 2.0 Tara virome datasets, encompassing polar stations, deep-ocean data, and Southern Ocean viromes. Viral protein sharing networks, using over 5000 representative viruses, were employed to assess the uniqueness of the RIS viroplankton. Host prediction utilized a combination of *in silico* approaches such as CRISPR spacer-protospacer matching and tRNA matching. Viral abundance and activity were estimated via metagenomic and metatranscriptomic fragment recruitment. Auxiliary metabolic genes (AMGs) were identified by comparing viral proteins against those from host MAGs and SAGs using BLASTP. Three-dimensional protein structure prediction was used for selected AMGs to compare the viral AMGs with those from potential hosts. Hydrophobicity indices of proteins were calculated using Protscale.

The study revealed a predominantly novel viral community beneath the RIS, with approximately 90% of assembled viral contigs belonging to Caudoviricetes (dsDNA viruses). A significant portion (around 50%) of the viral genes showed no homology to existing databases, highlighting the endemism and novelty of the RIS viroplankton. Network analysis confirmed the unique nature of many of these viruses, with a significant proportion appearing as singletons or outliers. The low abundance of integrase genes (<3%) did not support the predominance of the PtW hypothesis, consistent with the low-productivity habitat under the RIS. The study identified host-virus pairs, revealing active infection of key primary producers such as *Thioglobus* spp. and *Nitrosopumilus* spp., supporting a 'kill-the-winner' dynamic. Analysis of viral transcripts showed that some of the most highly transcribed viruses infected slow-growing microbes like *Pelagibacter* spp., potentially reflecting viral persistence in the low-temperature environment. A significant finding was the identification of various AMGs in the RIS viruses, notably those involved in iron-sulfur cluster protein synthesis (putatively boosting sulfur metabolism in chemolithoautotrophs), phosphorus acquisition (*PhoH* and acid phosphatase genes), and nitrogen metabolism (glutamine synthetase).

The findings challenge the existing understanding of viral ecology in low-productivity environments. The dominance of novel, endemic viruses and the lack of support for the PtW hypothesis suggest adaptation to the unique conditions under the RIS. The 'kill-the-winner' dynamic observed, along with the identification of AMGs, highlights the active role of viruses in shaping microbial community composition and function. The presence of AMGs involved in nutrient acquisition underscores the potential impact of viruses on biogeochemical cycles, not just as agents of mortality, but also as facilitators of host metabolism. The relatively high transcriptional activity of viruses infecting slow-growing bacteria suggests that viral activity is not solely coupled to high host growth rates in this environment, indicating a possible extension of viral persistence and influence. The discovery of novel viruses with AMGs has implications for our understanding of global nutrient cycling, particularly sulfur, phosphorus, and nitrogen, and the functioning of this previously poorly understood ecosystem.

This study provides a comprehensive characterization of the viral community beneath the RIS, revealing a unique and largely unexplored diversity of viruses with significant potential impact on biogeochemical cycles. The findings highlight the importance of considering viral activity and metabolic influences within these critical polar ecosystems. Future research should focus on culturing these novel viruses and experimentally validating the proposed roles of their AMGs in host metabolism and nutrient cycling. Furthermore, incorporating viral processes into models of the Antarctic ice shelf ecosystem and their effects on global nutrient cycles is needed.

The study's reliance on previously generated data limited the ability to directly assess the free-living viral community. The conservative approach used to identify bona fide viral genomes might have resulted in the underestimation of certain viral groups, particularly RNA viruses. The in silico predictions of host-virus interactions and AMG functions require experimental validation.