Microbiomes play a crucial role in the physiology of fish across various environments. Fish microbiomes are shaped by numerous environmental (i.e., water physicochemical parameters) and host (i.e., developmental stage, genetics, etc.) factors. While many studies have begun to explore microbiomes of rainbow trout, the exact metabolic functions of these microbiomes and their interactions with the host and its environment remain unclear. The two common methods used to study microbiomes (i.e., 16S rRNA gene sequencing and shotgun metagenomic sequencing) have their own set of limitations. 16S rRNA gene sequencing, for instance, is limited to identifying bacterial constituents, suffers from phylogenetic biases in PCR amplification, and provides no information of potential microbiome functions. Shotgun metagenomics, on the other hand, is relatively more expensive, computationally intensive, and often requires much deeper sequence to achieve the same level of characterization achieved by 16S rRNA gene studies.
We conducted a large-scale microbiome survey of commercial trout farms in Idaho using 16S rRNA sequencing data (presented here by Overturf). However, because of the limitations posed by 16S rRNA sequencing, a concurrent study, presented here, utilized shotgun metagenomics to evaluate trans-kingdom (bacteria, virus, fungi, etc.) phylogeny and metabolic functional potential of environmental (i.e., water, raceway biofilm, and diets) and host (i.e., fish gills, skin, and gut mucosa) microbiomes.
Raw samples from our large scale on-farm 16S rRNA study were pooled from 4th use rearing units for metagenomic evaluation of six discrete sample-types (i.e., raceway biofilm, diet, water, fish gill, skin, and gut mucosa). Samples were selected for metagenomic analyses according to DNA yield and good 16S rRNA results. Pooled DNA for each sample-type was shotgun-sequenced using Illumina technology. Reads were pre-processed using Trimmomatic and BBTools, taxonomic classification was determined using Kaiju and Kraken2. Metagenomic assemblies, binning, and annotations were conducted using the KBase platform and various command-line tools.
The level of host-contamination varied across sample types. Taxonomy results show a variation in the microbial composition of each sample type, although Protobacteria was the most dominant phylum across sample-types (Fig. 1). Further results presented will include methodological comparisons of bioinformatic steps, a deeper comparison of phylogeny by sample-type and technique (16SrRNA vs. shotgun) and insights on the metabolic functional potential of these on-farm environmental and host-associated microbiomes.