Atlantic salmon (Salmo salar), a key species in marine aquaculture, has experienced a production drop of over 35% since 2000, partly due to decreased embryo survival rates. Currently, selective breeding programs use metrics that are mostly based on growth related traits, but these traits do not correlate well with reproductive performance. We are developing measurement techniques to assess broodstock quality through the analysis of biofluids (plasma, ovarian fluid, and skin mucus) collected from Atlantic salmon female broodstock with the objective of identifying a suite of metabolic markers that correlate with reproductive success by comparing metabolite profiles in high performers (>70% eye-up rate eggs) and low performers (<70% eye-up rate eggs) using NMR metabolomics in combination with machine learning.
Plasma, ovarian fluid and skin mucus were collected in collaboration with USDA and the University of Maine. Two cohorts were sampled during spawning in 2021 and 2022. Plasma and ovarian fluid were filtered using centrifugal filters (3 kDa MWCO). Mucus samples collected on filter paper, were extracted with 70% methanol. NMR spectra (Fig. 1) were acquired at 298K on a Bruker Avance II 700 MHz spectrometer using both 1D and 2D NMR experiments (1H NOESY, and 1H,13C-HSQC). Spectra were binned, normalized to total spectral area, and scaled (Pareto) prior to multivariate analysis (PCA and PLS-DA). Metabolite identification was performed by comparison of experimental chemical shifts to reference values in available metabolomics databases (Chenomx, HMDB and BMRB), and an in-house library.
Our preliminary results from metabolomics analysis of 2022 data show that among the three different biofluids analyzed in this study, ovarian fluid NMR metabolite profiles showed better clustering according to egg eye-up rates. Since ovarian fluid is the obvious reproductive matrix, machine learning approaches will be used to identify correlating metabolic features in other biofluids like plasma and mucus to evaluate the fish metabolome for reproductive fitness. Skin mucus is of particular interest since it constitutes a more readily accessible and non-invasive biological matrix. Results from this study will guide the development of robust molecular tools for efficient broodstock selection, with the potential for early culling of low-yield broodstock, thereby enhancing reproductive success, and improving environmental sustainability.