REVOLUTIONIZING FISH EGG QUALITY ASSESSMENT: RAPID FATTY ACID DETECTION WITH RAMAN SPECTROSCOPY
Egg quality in aquaculture is often judged by physical traits (egg diameter, symmetry), but biochemical composition, especially n3 fatty acids, more directly predicts developmental potential. DHA, EPA and ARA are linked to higher hatching success and better egg performance. While GCMS provides accurate fattyacid profiles, it is slow, costly and laborintensive.
We evaluated Raman spectroscopy as a rapid, lowcost alternative to estimate DHA and EPA in fish eggs. Raman captures a nearinstant chemical fingerprint from intact samples and is sensitive to unsaturated lipids. We paired Raman spectra with GCMS reference data from 90 spawn samples collected over seven years from California yellowtail (Seriola spp.) and Atlantic cod (Gadus morhua). After preprocessing, a PLS regression model was trained to predict DHA and EPA from the spectra; model tuning and validation used crossvalidation and leaveseasonout tests to assess generalizability.
The model showed strong performance: R2CV = 0.91 and RMSECV = 2.5% (see Fig. 1, predicted vs. actual DHA+EPA). Figure 1 also highlights that Seriola samples span a wide quality range while cod samples cluster tightly, illustrating model behavior across different sample distributions.
These results demonstrate Raman spectroscopy as a rapid, portable, and costeffective tool for egg quality screening. Deploying Ramanbased monitoring in hatcheries could enable nearrealtime assessment of reproductive performance and help ensure consistent production of robust offspring across multiple species.
Figure 1: Partial least squares (PLS) regression plot of actual versus predicted EPA+DHA values using Raman spectra in California yellowtail and Atlantic cod eggs