NIRS analysis to assess the nutritional value of two corn hybrids dried at different temperatures: a comparative analysis between laboratories

Near infrared reflectance spectroscopy (NIRS) is nowadays a common tool to analyze the nutritional content of feedstuffs. Corn hybrids and drying temperatures could affect the nutritional value of the grain. It is important to determine if these factors are detected by NIR. Two corn hybrids (Dekalb 68-05 and Dekalb 65-20) varying in kernel hardness (average and hard respectively) were dried at three temperatures (35, 80, and 120°C) to investigate the effects on its nutritional content. Subsequently, 5 samples of whole and grounded corn per treatment were collected and read in a FOSS (DS2500) NIR. The spectrum of each sample was sent to 5 commercial companies (AB Vista, Adisseo, Cargill, DSM, and Trouw Nutrition) to use their respective calibration models. Results from each commercial company were codified as laboratories A to E independently of the alphabetical sequence presented above. Treatments resulted from a 2 x 3 factorial arrangement of two corn kernel hardness and three drying conditions. Data were analyzed in a completely randomized design and mean separation with Tukey’s test. Whole grain samples of non-dried grain showed that the corn with hard kernel had greater (P < 0.05) moisture in all models, vitreousness (A: 89.17 vs. 83.44%), starch (A: 83.57 vs. 78.71%), and fat (B: 4.28 vs. 3.88%) compared with the corn with average hardness. In contrast, no differences (P > 0.05) were observed in protein with these samples. However, once the corn hybrids were dried, results indicated that both grain hybrids dried had the lowest (P < 0.01) protein at 80°C and higher (P < 0.01) fat content at 35°C in grounded samples. At 35°C higher (P < 0.01) content of Lys, Met+Cys, and Thr were observed in corn with average hardness. Fatty acid content was greater (P < 0.001) in corn with hard kernel compared with the corn of average hardness and effects of drying temperatures also were detected. Hard kernel corn had less starch (P < 0.001) than corn with average hardness when dried at 80 and 120°C in all models. These results could affect the predicted energy for feed formulation. In conclusion, NIRS was able to detect differences between corn hybrids dried at different temperatures independently of the calibration model, but there was variation in results among models.

Key Words: NIRS, corn hybrids, nutritional value, drying temperature.