<br/>The increased demand for marine oil dietary supplements requires rapid and accurate screening to verify the accuracy of label declarations. Simple and rapid procedures based on spectra from a portable Fourier-transform infrared (FTIR) device and a benchtop FT-near infrared spectrometer were evaluated for predicting the fatty acid (FA) concentrations in a representative set of marine oil products (n=95). Additionally, two chemometric calibration approaches for each spectrometer were tested. Separate calibration models using partial least squares regression (PLS-R) and support vector regression (SV-R) were developed to correlate the FA concentrations of the marine oils with spectroscopic responses. The developed models were successfully validated by cross-validation, testing an independent sample set, and comparing to a commercially available fish oil standard reference material (SRM). Both PLS-R and SV-R model predictions showed good linear correlations with a reference gas chromatographic method (R<sup>2</sup> &gt; 0.91) and excellent predictive qualities for eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA). When compared, SV-R outperformed PLS-R as shown in improved accuracies as well as prediction qualities. For the six FA/FA class models, the prediction errors for the SV-R models developed for the FTIR spectral data were decreased by as much as 20-58% compared to the corresponding PLS-R models. Similarly, the prediction performance of FT-NIR models were also improved (i.e. RMSEP values were decreased by 25-58%). Further, the SV-R models yielded better agreement with the NIST certified/reference values for the prediction of EPA and DHA concentrations in three SRM samples. This higher performance of SV-R can be attributed to its ability to better quantify the non-linear relationships in marine oil samples. This study confirmed that the simple, fast, and nondestructive quantitative procedures based on PLS-R and SV-R coupled with ATR-FTIR and FT-NIR spectroscopies can be successfully applied for the rapid screening of marine oil products to verify the accuracy of label declarations.<br/>

Linear and non-linear calibration approaches for the rapid quantification of marine oil omega-3 supplements using vibrational spectroscopy

2019 Fall National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.