Non-destructive prediction of total phenolics and antioxidants in hulled and naked oat genotypes with near-infrared spectroscopy

Oats have gained significant attention due to their multiple health benefits. In the present research, hulled and naked oat samples were collected from various parts of China, Canada, and the United States. These samples were explored for total phenolic content (TPC), total flavonoid content (TFC), and antioxidant potential as assessed by DPPH assay. Near-infrared spectroscopy (NIRS) was employed to establish calibration models for non-destructive prediction of TPC, TFC and DPPH values. The principal component analysis (PCA) was employed for the second derivative of NIR spectra of oats resulting in satisfactory discrimination of naked and hulled oats. NIR spectra in conjunction with partial least square regression (PLSR) were employed to establish calibration models. The hulled and naked oats are a rich source of TPC and exhibit potent antioxidant activity. PCA of spectral data resulted in an efficient classification of naked and hulled oats. NIR spectra along with PLSR were employed to develop calibration models (R-2 = 0.65 and 0.67) for moderate prediction for TPC in oat samples. Furthermore, more samples from different geographical locations with significant variations in their chemical composition need to be included in the sample pool to develop robust calibration models for strong prediction of all attributes.

Automated summary

Oats from different parts of China, Canada, and the United States were investigated for their phenolic and flavonoid content as well as antioxidant potential. NIR spectra combined with PLSR were used to develop calibration models for the non-destructive prediction of these attributes. However, more diverse samples are needed to establish robust calibration models for accurate prediction.