Genome-WIDe Asssociation study of soybean meal tolerance in Atlantic salmon  

Tiago Hori*, André Dumas, Jason Stannard, Debbie Plouffe and John Buchanan
The Center for Aquaculture Technologies Canada
20 Hope Street
Souris (PE) C0A 2B0, Canada
thori@aquatechcenter.com

The utilization of Fish Meal (FM) as a protein source in the Aquaculture industry has been shown not to be sustainable. Soybean meal (SBM) stands as viable replacement for fishmeal in aquaculture because of the relatively high protein content, amino acid profile and digestibility, as well as its global availability and low cost. However, SBM inclusion in salmon feed is hindered by poor performance and negative physiological impacts (e.g. enteritis). Studies in trout have shown that there is significant individual and family variation in the ability to use SBM derived protein in salmonids, and that this trait has moderately high heritability. Therefore, there is potential for the selection of lineages with higher tolerance to SBM inclusion. Marker assisted selection, (MAS) using genetic markers such as single nucleotide polymorphisms (SNPs) associated with traits of interest, can be used to compliment traditional breeding and reduce the time required to achieve genetic gains.

Fin samples from families presenting poor and good growth when fed a high SBM diet were selected and DNA was extracted from fins. Genotyping-by-Sequencing (GBS) was used to obtain genotypes for a total of 291 animals. In total, over 90,000 putative markers were identified, and 46,333 SNPs were selected for GWAS analysis (MAFs > 0.01 and rate of genotyping > 85%). Missing data was imputed using the LDKNNimp algorithm was implemented in the TASSEL v5 pipeline. Association analysis was performed using a weighted mixed linear model with kinship (measured by centered IBS) and population structure (measured by PCA). The thermal coefficient growth (TGC) data was fitted to the model and the effects of each SNPs estimated one by on using both a dominant and an additive model. The Bonferroni approach was used to control for family-wise error (FWER) (i.e. the obtained p-value was divided by the number of tests performed - 0.05/46,333 - p-vaule cutoff of 1.08 X 10-6).

Using this approach, seven SNPs (in 5 different chromosomes) were identified as being significantly correlated with TGC after FWER correction. These markers explain had an average R-squared of 0.09 and explain roughly 9% of the estimated heritability of TGC of Atlantic salmon when fed diets containing high levels of soybean meal. The markers, if validates, can be added to breeding programs and facilitate higher inclusions of soybean meal in commercial diets for Atlantic salmon.