Aquaculture is of major and increasing importance to global food security, particularly in Low Income, Food Deficit Countries (LIFDCs), where it also serves as a significant contribution to poverty alleviation. Disease is widely acknowledged as the prominent bottleneck to achieving global food security and poverty alleviation targets relating to aquaculture, with annual losses exceeding >$6bn (Food and Agriculture Organization 2014). Outbreaks of disease caused by endemic and emerging pathogens impact directly on farmer income and their nutritional security. Avoidance of yield-limiting disease outbreaks is a fundamental requirement for future growth and resilience of aquaculture in LIFDCs. Advances in molecular techniques coupled with next-generation sequencing have provided a step-change in understanding the role of host-associated bacteria, archaea, protists and viruses (the microbiome) in host homeostasis. Shifts in microbiome communities under stressful conditions can contribute to disease states. However, the role of microbiomes in the emergence of diseases in aquaculture, where stressors include feeding, antibiotic and disinfectant use and over-stocking, is poorly studied.
Here our study presents an evaluation of the microbiomes (bacteria and viruses) associated with tilapia and their pond environments in aquaculture, using 16S rRNA community profiling techniques and viral amplicon sequencing. Samples investigated in this project were collected from Malawi tilapia fish farms; their skin community composition and diversity were examined across geographical scales. The high variability observed of the microbial communities in small geographic regions, showed that future sampling to detect shifts due to dysbiosis will require time-resolved sampling of ponds under study. Nanopore sequencing of full length 16S rRNA genes, using MinION, allowed us to examine the microbial communities at higher taxonomic resolution than short read sequencing techniques. Its success lays the foundation for in-situ microbial profiling of aquaculture ponds for disease, and offers independence to farmers to monitor their own ponds. Successful amplification of the T4-like Myoviridae phylogenetic markers from one rearing water sample was achieved, although the required degeneracy of the primers inhibited multiplexing. Therefore, our findings suggest that inclusion of bacteriophages in microbiome studies is better served using shotgun metagenomic methods, rather than amplicon sequencing.

Finally, we investigated the use of skin swabbing as an alternative to bucket incubations to minimise animal stress when categorising the fish skin microbiome. Skin swabbing successfully captured similar microbial communities in comparison to bucket incubations, with greater diversity and variance between fish.