Near-infrared spectroscopy (NIRS) evaluation and regional analysis of Chinese faba bean (Vicia faba L.)

To analyze the nutritional composition of faba bean (Vicia faba L.) seed, estimation models were developed for protein, starch, oil, and total polyphenol using near infrared spectroscopy (NIRS). Two hundred and forty-four samples from twelve producing regions were measured in both milled powder and intact seed forms. Partial least squares (PLS) regression was applied for model development. The model based on ground seed powder was generally superior to that based on the intact seed. The optimal seed powder-based models for protein, starch, and total polyphenol had coefficients of correlation (r2) of 0.97, 0.93 and 0.89, respectively. The relationship between nutrient contents and twelve producing areas was determined by two-step cluster analysis. Three distinct groupings were obtained with region-constituent features, i.e., Group 1 of high oil, Group 2 of high protein, and Group 3 of high starch as well as total polyphenol. The clustering accuracy was 79.5%. Moreover, the nutrition contents were affected by seeding date, longitude, latitude, and altitude of plant location. Cluster analysis revealed that the differences in the seed were strongly influenced by geographical factors.