Strategies to determine lactose in cow milk by mid infrared spectroscopy

The lactose-free dairy sector has been growing considerably, requiring rapid and accurate methodologies for lactose determination. The present work aimed to explore spectroscopy and statistic strategies to estimate the lactose content in cow milk using mid infrared spectroscopy (MIRS) and chemometric tools. Firstly, regular and lactose-free milk discrimination was successfully performed using the spectral range of 935-1200 cm-1 along with Partial Least Squares Discriminant Analysis (PLS-DA). Secondly, to estimate the percentage of lactose in lactose-free milk, calibration models were developed by Partial Least Squares (PLS), Multiple Linear Regression (MLR), and Least Squares Regression (MQ) with and without spectral transformation. The three methods proved to be efficient, with the best performance obtained by the PLS model using Multiplicative Scatter Correction from 935 to 1200 cm-1, with low RMSE values and R2 > 0.99 for calibration, cross and external validation. Furthermore, high-performance liquid chromatography (HPLC) was used to attest the good predictive ability of lactose content in lactose-free milk by MIRS-PLS. Finally, the models were used to monitor the lactose content during the enzymatic hydrolysis, showing the applicability and efficiency of the proposed method. Therefore, MIRS associated with chemometric tools constitutes a method with a high capacity to discriminate between regular and lactose-free milk, as well as to predict lactose content in both milk samples.