Phenomics

Raman-based prediction of fillet fatty acid composition and genetic architecture in Atlantic salmon (Salmo salar)

---

Abstract

Atlantic salmon (Salmo salar) is an economically important aquaculture species and a major food source of omega-3 long-chain polyunsaturated fatty acids (n-3 LC PUFA), such as eicosapentaenoic (EPA; C20:5n-3) and docosahexaenoic acid (DHA; C22:6n-3). The total lipid content and the fatty acid (FA) composition of salmon fillets are heritable and influence flesh quality, nutritional value, and marketability. Traditional laboratory-based methods, such as gas chromatography (GC), for quantifying FA composition are accurate but also costly, time-consuming, and destructive to samples, limiting their application in large-scale selective breeding programs. Raman spectroscopy offers a rapid, cost-effective, and non-destructive alternative for high-throughput phenotyping. The objective of this study was to evaluate the feasibility of using Raman spectroscopy to predict omega-3 FA contents of Atlantic salmon fillets. Secondly, to estimate the heritability, genetic correlations, and genomic architecture of Raman-predicted FA traits. A total of 834 fillet samples of Atlantic salmon reared under saltwater conditions until 1.3 kg originating from 30 full-sib families from the AquaGen strain were scanned using Raman spectroscopy. Five traits were evaluated: total polyunsaturated fatty acids (PUFA), EPA, DHA, α-linolenic acid (ALA; C18:3n-3), and the sum of EPA and DHA. A Partial Least Squares Regression (PLSR) model was applied for FA prediction, trained using GC reference data as calibration. Genetic parameters were estimated using linear mixed models with genomic relationship matrices based on 43,644 single-nucleotide polymorphisms (SNPs). Genome-wide association studies (GWAS) were performed using GCTA software with Bonferroni correction for genome-wide significance. The PLSR model prediction using Raman spectroscopy showed moderate to good predictive ability, with R-squared (R2) ranging from 0.55 to 0.81, and Root Mean Squared Error (RMSE) ranging from 0.08 to 0.29. Heritability estimates ranged widely from 0.07 ± 0.03 (EPA) to 0.42 ± 0.07 (DHA) and 0.44 ± 0.06 (PUFA), indicating their potential as selection traits. The genetic correlations (RG) were moderate to high between related FA, including 0.77 between PUFA and EPA, and 0.97 between EPA and DHA. GWAS results confirmed a highly polygenic genetic architecture for all traits, with no single SNP explaining a large proportion of variance. Raman spectroscopy showed potential for a high-throughput phenotypic tool for omega-3 FA in Atlantic salmon fillet under commercial saltwater conditions. These findings provide more concrete evidence for the feasibility of using rapid spectroscopic measurements of fatty acid profiles in salmonid breeding programs.

Keywords: 2026

How to Cite:

Park, J., Difford, G., Ahmad, A., Jacob, N., Ødegård, J., Dvergedal, H., Klemetsdal, G., Wold, J., Berg, P., Lintvedt, T. & Afseth, N., (2026) “Raman-based prediction of fillet fatty acid composition and genetic architecture in Atlantic salmon (Salmo salar)”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2285433. doi: https://doi.org/10.31274/wcgalp.23705

Rights: 1