Introduction
Fish transport is an essential yet stressful practice in aquaculture. Acute transport-induced stress can compromise both specific and non-specific immunity, disrupt osmoregulation, and increase disease susceptibility, potentially leading to delayed mortality. Monitoring fish welfare physiologically remains challenging, as current methods rely on invasive procedures such as blood sampling, tissue collection, or surgical implantation of biosensors. Moreover, the established indicators are often detected too late for timely intervention. The skin and mucus of teleost fish are central to innate and adaptive immunity and homeostasis. Given the feasibility of minimally invasive or non-invasive sampling, they represent promising matrices for biomarker discovery enabling earlier and less intrusive health monitoring in farmed species.
Materials and methods
To investigate the molecular signatures following acute stress, three key European aquaculture species: Atlantic salmon (Salmo salar), European seabass (Dicentrarchus labrax), and rainbow trout (Oncorhynchus mykiss) were subjected to a transport challenge. The transportation trials were designed to simulate current transport conditions specific to each species. Skin tissue and mucus samples were collected at three time points: before transport, immediately after stress exposure, and following a 24-hour recovery period. Bulk polyA RNA sequencing was conducted on skin samples, while small RNA sequencing and proteomics were performed on skin mucus. Sequencing data were processed using nf-core pipelines for RNA-seq (nf-core/rnaseq) and small RNA-seq (nf-core/smrnaseq), while proteomic data were analysed using the FragPipe pipeline. Differential expression analysis (DEA) was performed using edgeR (transcriptomic) and FragPipeAnalystR (proteomic) R packages. Additionally, unsupervised machine learning based on Multi-Omics Factor Analysis (MOFA+) was performed. The results of DEA and MOFA were compared and similarities highlighted.
Results and Discussion
Differential expression analysis, complemented by MOFA multiomics integration, revealed species-specific stress responses differing in both timing and magnitude. Enrichment analyses and shared element comparisons further highlighted the negative physiological impact of transportation stress across species. The acute transcriptomic response, based on the number of differentially expressed genes immediately following stress, was markedly stronger in Atlantic salmon (3,134 DEGs) and rainbow trout (1,241 DEGs) compared to European seabass (309 DEGs). Conversely, European seabass exhibited a stronger and more rapid skin mucus proteomic response, being the only species with differentially abundant proteins detected at the post-stress timepoint. Cross-species comparison identified a conserved set of 42 genes and 3 miRNAs while no protein was found to be commonly differentially abundant across all three species. To filter out candidate biomarkers, elements consistently identified across at least four subgroups defined by species, pairwise comparisons, and omic layers, were highlighted. The 17 highlighted biomarker candidates are primarily associated with coagulation, extracellular matrix remodelling, wound healing, and immune response which was also reflected in the observed commonly activated pathways, while pathways associated with biosynthesis and metabolism were inhibited. Among the candidates, arginase (ARG2), ferroxidase (F5), and papilin a (PAPLNA) were the most consistently represented across subgroups. Independently, MOFA integration identified factors that separated samples by condition in all three species. While several candidate biomarkers were recovered across species and omic layers, none were consistently top-weighted across all three species in the factors relevant to stress, emphasising the complexity of the stress response and the need for multiple complementary discovery approaches. Future analyses incorporating supervised machine learning and expanded datasets will further refine biomarker prioritisation and improve fish welfare assessment.
Acknowledgment
Work Co-funded by UKRI and by the European Union's Horizon Europe research and innovation programme (GA No. 101084651 - project IGNITION). TB received funds from FCT through grant 2023.04651.BDANA.