Deep-sea shipwrecks represent island-like ecosystems for marine microbiomes

Microbial biogeography, the non-random distribution of microorganisms across space and time, is a complex field of study. While considerable progress has been made in understanding microbial biogeography in various environments, the deep ocean remains relatively unexplored due to access limitations. Existing research suggests that both historical events (e.g., the evolution of microbiomes in isolated habitats) and contemporary environmental factors (e.g., local physical and geochemical characteristics) shape microbial biogeographic patterns. Key drivers of microbial biogeography in aquatic environments include pelagic dispersal and environmental selection (physical and chemical cues). However, much of the research focuses on natural features, with built features like shipwrecks receiving limited attention. This study addresses this gap by focusing on the influence of a deep-sea shipwreck on the surrounding microbiome. The researchers hypothesize that shipwrecks act as biodiversity hotspots, impacting the distribution and composition of marine microbiomes in a manner analogous to islands in terrestrial ecosystems. The *Anona* shipwreck in the Gulf of Mexico serves as a case study to examine the scale and extent of this impact.

Previous studies have extensively explored environmental factors influencing microbial community composition and assembly. A review by Hanson et al. [3] highlighted the significant influence of physical and chemical factors, including hydrodynamics (density, salinity, temperature, bathymetry, and circulation) on the dispersal and biodiversity of aquatic organisms, including microorganisms. The review showed that a correlation exists between microbial composition and habitat features in 92% of studies, reinforcing the importance of habitat features in shaping microbial biogeographic patterns. In the deep sea, natural features such as methane seeps, hydrothermal vents, and seamounts are known to support unique microbial communities. However, the influence of artificial structures, such as shipwrecks, on deep-sea microbial biogeography remains largely uncharted.

Sediment samples were collected from around the *Anona* shipwreck in the Gulf of Mexico during seven expeditions using various methods including a deep multi-corer, ROVs, and push cores. Samples were collected at distances ranging from 2 to 200 m from the wreck along four transects. Visual surveys were conducted using sonar to ensure sampling locations were free from shipwreck debris. Samples were also collected from two additional shipwrecks, Halo and Alcoa Puritan, but their use was limited to a machine learning exercise to identify ASVs associated with proximity to large steel-hulled shipwrecks. Genomic DNA was extracted from the sediment samples, and the V6–V8 variable regions of the bacterial 16S rRNA gene were amplified and sequenced using Illumina MiSeq platform. Bioinformatics analyses were performed using QIIME2, including DADA2 for quality control and taxonomic identification using SILVA version 132 reference. Statistical analyses were conducted using PRIMER v. 6.1.13, including non-metric multidimensional scaling (nMDS), principal coordinates analysis (PCoA), hierarchical clustering, ANOSIM, and SIMPER to analyze community structure and identify key taxa driving differences between sample groups. To assess island effects, diversity, taxa, and evenness extinction plots were created. A Random Forest Regressor was employed for machine learning analysis to predict ASVs associated with proximity to the shipwrecks.

A total of 138 sediment samples were collected around the *Anona* shipwreck, yielding approximately 700,000 sequences after quality control. PERMANOVA analysis revealed significant effects of both distance from the shipwreck and sediment depth on microbiome community structure. nMDS ordination showed distinct bacterial groupings based primarily on distance from *Anona*. Samples from 0–200 m formed four clusters, with the largest cluster encompassing the entire distance range. Samples collected beyond 300 m were statistically different from those within 200 m. Approximately 17% of ASVs were classified only at the phylum level as Proteobacteria. Gammaproteobacteria, Phycisphaerae, Deltaproteobacteria, and Alphasubproteobacteria were among the dominant phyla. SIMPER analysis identified specific taxa contributing to the differences between sample clusters. The study demonstrated a clear island effect, with increased microbiome richness and diversity observed in closer proximity to the shipwreck, and a transition zone extending up to 200m away. This extended transition zone is significantly larger than those observed around natural deep-sea habitats.

The findings strongly support the hypothesis that deep-sea shipwrecks act as 'islands' of biodiversity for marine microbiomes. The observed 'island effect', characterized by increased richness and diversity near the shipwreck and a substantial transition zone, highlights the significant ecological role of these artificial structures. The extended spatial influence of the *Anona* shipwreck, compared to natural habitats, suggests that human-made structures may exert a disproportionately large impact on the surrounding benthic environment. These results contribute to a growing understanding of how human activities, even those occurring decades or centuries ago, can shape the biogeography of deep-sea ecosystems. The study underscores the importance of considering both natural and artificial structures when investigating microbial biogeography in the deep sea.

This study provides compelling evidence for the significant influence of deep-sea shipwrecks on the surrounding marine microbiomes. The observed island effect and extended transition zone highlight the ecological importance of these artificial habitats. Future research could investigate the long-term effects of shipwreck presence, the role of specific shipwreck characteristics (e.g., material composition, size, age) on microbiome structure, and the broader implications for deep-sea biodiversity and ecosystem function. Further comparative studies incorporating various shipwreck types and environmental settings would strengthen our understanding of the ecological impacts of human-made structures in the deep ocean.

The study focused on a single shipwreck in the Gulf of Mexico, limiting the generalizability of the findings. While additional shipwrecks were sampled, their limited sampling design restricted their use primarily for machine learning applications. Future research should consider a broader range of shipwrecks to validate and expand upon these findings. The study also concentrated on bacterial communities, and future work might explore other microbial groups like archaea and fungi to understand the overall impact of shipwrecks on the deep sea microbiome. Further research to assess the functional roles of these microbial communities and their interactions with the surrounding environment would enrich this work.