What makes a good egg? In aquaculture, egg quality is typically determined through physical metrics such as egg and oil diameter or egg symmetry. While these are reliable indicators, they are not sufficient to fully capture egg quality or predict early larval performance. The biochemical composition of eggs which include essential nutrients such as lipids is critical in fueling embryogenesis and can serve as powerful indicators of egg quality. However, analyses for these nutrients require costly analytical equipment making them inaccessible and impractical for on-farm applications. Our research aims to overcome this challenge by developing a rapid, on-site method to assess egg quality using chemometric modeling coupled with Raman spectroscopy.
Raman spectra, paired with corresponding fatty acid data obtained via GC-MS, were collected and analyzed from 90 unique spawns of California yellowtail (Seriola dorsalis) collected over seven years (2016 – 2024). Eggs collected had a wide range of fatty acid composition to allow for a more comprehensive model. Using this dataset, a partial least squares (PLS) model was developed for two essential omega-3 fatty acids, DHA and EPA. The robustness of the model was further tested and validated using cross-validation, where spawns from different years were held out to provide a realistic assessment of the model’s predictive power across natural variation in egg quality over multiple spawning seasons.
The model showed strong predictive power and high accuracy (Figure 1; R²CV= 0.91, RMSECV= 2.5%), demonstrating the viability of Raman spectroscopy as a novel tool to rapidly access biochemical information in fish eggs. This approach could transform broodstock management approaches and open new opportunities for research in broodstock nutrition and egg quality. Future work for this project will focus on developing a multi-species model and extending predictions to other nutrients such as carotenoids.