The employment of Vis-NIR and FTIR spectroscopy in combination with chemometrics for the authentication of Patin fish oil

• Vis-NIR and FTIR enable rapid detection of PFO adulteration • LDA classifies authentic vs adulterated PFO with ≥ 93% accuracy • PLSR predicts palm oil levels with R² ≥ 0.97 • Spectroscopy offers a rapid and green authentication method Patin fish oil (PFO), extracted from Pangasius micronemus , is a valuable source of omega-3 polyunsaturated fatty acids (PUFAs). However, its authenticity is compromised through adulteration with lower-cost vegetable oils, particularly palm oil. This study developed rapid, reliable non-destructive models using Visible-Near-Infrared (Vis-NIR, 600–2500 nm) and attenuated total reflectance Fourier Transform Infrared Spectroscopy (ATR–FTIR, 4000–600 cm⁻¹), combined with chemometrics. A total of 175 samples were analyzed, comprising authentic PFO, crude PFO (CPFO), and adulterated mixtures with palm oil at concentrations ranging from 1% to 99% (v/v), prepared from three different batches of oil to account for natural compositional variations. Each sample was measured in triplicate, yielding 525 spectra per technique. Principal component analysis (PCA) was used for exploratory analysis, while linear discriminant analysis (LDA) and partial least squares discriminant analysis (PLS–DA) were used for classification. Partial least squares regression (PLSR) was used for quantitative predictions. The dataset was split into calibration (75%, n=131) and validation (25%, n=44) sets using stratified sampling. Using optimized pre-processing standard normal variate (SNV for Vis-NIR and first derivative for FTIR), LDA achieved ≥93% accuracy with 100 % sensitivity for authentic samples. PLSR models demonstrated strong predictive performance (Vis-NIR: RMSEP = 4.78%, R² = 0.978; FTIR: RMSEP = 5.1%, R² = 0.98). These findings highlight the complementary capabilities of Vis–NIR and FTIR spectroscopy for rapid and green authentication of fish oil.