Florian Baletaud

Environmental DNA (eDNA) metabarcoding produces sparse read-count matrices routinely binarised to presence/absence to mitigate amplification, copy-number and shedding biases. We argue that binary eDNA data are sufficient to recover detailed community structure, sparse indicator panels and habitat-specific co-occurrence sub-networks — given the right analytical pipeline. Yet integrative frameworks bridging taxon-level and community-level structure remain missing for macro-organism eDNA. We propose a reproducible framework combining three scales: sparse machine-learning indicator-species identification (Predomics), consensus co-occurrence network inference (ScaleNet), and module-level enrichment via Gene Set Enrichment Analysis (GSEA). We applied it to 217 eDNA samples spanning eight habitats and three depth zones (shallow, <50 m; middle, 50–250 m; deep, 200–600 m) in New Caledonia, South Pacific. A sparse 55-MOTU indicator panel captured a proportion of beta-diversity variance comparable to the full 318-MOTU community. ScaleNet reconstructed a 261-node, 579-edge network with a core-periphery main component (shallow in the core, middle at the periphery) and a structurally isolated deep-zone subgraph. Fast-Greedy modularity partitioned it into 13 modules, 11 significantly enriched for depth zones or habitats by GSEA, each mapping onto a coherent trophic guild — from mesopelagic lanternfishes over deep seamounts to reef grazers across shallow microhabitats. Indicator taxa occupied structurally peripheral positions while connectivity hubs were non-indicator generalists, with reef-scraping parrotfish a functionally important exception. Sparse indicator panels efficiently capture zone discrimination while co-occurrence networks reveal community-level integration — together providing a scalable, interpretable framework for eDNA-based marine biodiversity monitoring.

Deep-sea demersal fisheries in the Pacific have strong commercial, cultural, and recreational value, especially snappers (Lutjanidae) which make the bulk of catches. Yet, managing these fisheries is challenging due to the scarcity of data. Stereo-Baited Remote Underwater Video Stations (BRUVS) can provide valuable quantitative information on fish stocks, but manually processing large amounts of videos is time-consuming and sometimes unrealistic. To address this issue, we used a Region-based Convolutional Neural Network (Faster R-CNN), a deep learning architecture to automatically detect, identify and count deep-water snappers in BRUVS. Videos were collected in New Caledonia (South Pacific) at depths ranging from 47 to 552 m. Using a dataset of 12,100 annotations from 11 deep-water snapper species observed in 6,364 images, we obtained good model performance for the 6 species with sufficient annotations (F-measures >0.7, up to 0.87). The correlation between automatic and manual estimates of fish MaxN abundance in videos was high (0.72 – 0.9), but the Faster R-CNN showed an underestimation bias at higher abundances. A semi-automatic protocol where our model supported manual observers in processing BRUVS footage improved performance with a correlation of 0.96 with manual counts and a perfect match (R=1) for some key species. This model can already assist manual observers to semi-automatically process BRUVS footage and will certainly improve when more training data will be available to decrease the rate of false negatives. This study further shows that the use of artificial intelligence in marine science is progressive but warranted for the future.

The regular and consistent monitoring of marine ecosystems and fish communities is becoming more and more crucial due to increasing human pressures. To this end, underwater camera technology has become a major tool to collect an important amount of marine data. As the size of the data collected outgrew the ability to process it, new means of automatic processing have been explored. Convolutional neural networks (CNNs) have been the most popular method for automatic underwater video analysis for the last few years. However, such algorithms are rather image-based and do not exploit the potential of video data. In this paper, we propose a method of coupling video tracking and CNN image analysis to perform a robust and accurate fish classification on deep sea videos and improve automatic classification accuracy. Our method fused CNNs and tracking methods, allowing us to detect 12% more individuals compared to CNN alone.

Seamounts are the least known ocean biome. Considered biodiversity hotspots, biomass oases, and refuges for megafauna, large gaps exist in their real diversity relative to other ecosystems like coral reefs. Using environmental DNA metabarcoding (eDNA) and baited video (BRUVS), we compared fish assemblages across five environments of different depths: coral reefs (15 m), shallow seamounts (50 m), continental slopes (150 m), intermediate seamounts (250 m), and deep seamounts (500 m). We modeled assemblages using 12 environmental variables and found depth to be the main driver of fish diversity and biomass, although other variables like human accessibility were important. Boosted Regression Trees (BRT) revealed a strong negative effect of depth on species richness, segregating coral reefs from deep-sea environments. Surprisingly, BRT showed a hump-shaped effect of depth on fish biomass, with significantly lower biomass on coral reefs than in shallowest deep-sea environments. Biomass of large predators like sharks was three times higher on shallow seamounts (50 m) than on coral reefs. The five studied environments showed quite distinct assemblages. However, species shared between coral reefs and deeper-sea environments were dominated by highly mobile large predators. Our results suggest that seamounts are no diversity hotspots for fish. However, we show that shallower seamounts form biomass oases and refuges for threatened megafauna, suggesting that priority should be given to their protection.

Coral reefs host the highest fish diversity on Earth despite covering less than 0.1% of the ocean’s seafloor. At the same time they are also extremely threatened. Data syntheses over decades of surveys estimate the total number of coral reef fishes to vary from 2,400 to 8,000 species distributed among roughly 100 families. But this diversity remains largely unknown. Here, we investigated how environmental DNA (eDNA) could describe the distribution of fish diversity in coral reefs. We generated 504,457,267 raw 12S ribosomal DNA (rDNA) sequence reads from 251 samples (2,693 PCR replicates) collected at 25 sites in 145 stations covering five regions across the Indian, Pacific and Atlantic Oceans. Bioinformatic analysis clustered these sequences into 2,160 molecular operational taxonomic units (MOTUs) corresponding to distinct species (Marques 2020) We compared our results, with visual census surveys from Reef Life Survey, on 2,813 transects in tropical regions. Our outcomes demonstrate the capacity of eDNA metabarcoding from water samples to reconstruct well-known biogeographic patterns of fish diversity on coral reefs, such as species richness gradients towards the coral triangle, and family proportion stability across sites (Bellwood and Hughes 2001). Additionally, eDNA survey data documented a higher fish species (16%) and family (50%) diversity than estimates obtained with underwater visual surveys carried out on 20 times more sites. MOTU richness per family retrieved with eDNA closely matched fish species richness within families recorded in visual census data. However, eDNA revealed higher richness of reef-associated species and species from adjacent habitats, of cryptobenthic or nocturnal species, but also of pelagic and wide-ranging species. eDNA survey data showed that fish diversity is characterized by spatially heterogeneous species assemblages among regions, with more dissimilarity among adjacent coral reefs than detected with visual survey data. Unlike visual surveys, eDNA metabarcoding revealed the same prevalence of rarity as expected under the neutral theory of biodiversity, suggesting the predominance of random processes and ecological equivalence within trophic groups at large scale to explain fish biodiversity patterns on coral reefs. Our study demonstrates how sequencing eDNA from water provides a rapid and effective approach to characterize and assess coral reef diversity across large spatial scales, thereby also uncovering hidden biodiversity patterns.