Evaluation of dietary fiber of Brazilian soybean (Glycine max) using near-infrared spectroscopy and chemometrics

• Ground soybeans varieties were analyzed by NIR without any further preparation.
• The variable importance for projection (VIP) was applied for selection variable.
• The statistical parameters of PLS calibration model for fiber showed good results.
• NIR can replace labor analysis of dietary fiber.

The dietary fiber analytical methods measure soluble and insoluble non-starch polysaccharides which include cellulose, hemicellulose, pectin, β-glucans, gums, and lignin. These assays are time-consuming and often produce a considerable amount of chemical waste. In this research, a methodology based on Fourier transform near-infrared spectroscopy (NIRS) in conjunction with multivariate calibration was developed for prediction of the total dietary fiber in soybean from Brazil (n = 80) using a traditional assay as reference. The total dietary fiber content of soybean samples using an enzymatic-gravimetric method varied between 10.64 g 100 g−1 and 19.50 g 100 g−1. The ground beans were directly analyzed by diffuse reflectance in the region between 10,000 and 4000 cm−1 and the variable importance for projection (VIP) was the method for variable selection, which reduced to 4500 variables. The best calibration results using partial least squares (PLS) regression satisfactorily allowed the determination of the concentrations of dietary fiber with acceptable coefficient of determination (R2 = 0.80) and low root mean square error of prediction (RMSEP = 0.86). It is confirmed that the NIRS technology is a method that may be useful to replace the traditional methods for routine analysis of dietary fiber in Brazilian soybean rapidly and accurately.