Files
Abstract
Because feed constitutes approximately 70% of the cost of broiler live production, developing diets that meet the needs of maintenance, growth, and production while minimizing nutrient excess, is one way to maximize profit and animal health. Protein, as a supply of digestible, essential amino acids, supports the rapid muscle growth of broilers and the prolific egg production of laying hens. Determining the digestible amino acid content of feedstuffs requires an animal bioassay and laboratory analyses, which are time consuming and expensive. In commercial poultry production, the results from these analyses are not available before the tested ingredient is fed. Near infrared reflectance spectroscopy is a rapid analysis technique involving calibration of a multi-purpose analyzer to detect the reflectance of near infrared light from a feed ingredient to predict the nutrient content of a sample. The current research strives to create a calibration curve for the prediction of digestible amino acid content for each of the 10 essential amino acids for poultry in feed ingredients used in poultry diets. Initially, a calibration curve for detection of each of the 10 essential amino acids was created and validated for use in any conventional or alternative feed ingredient. While the R2 values of the validation curves for the 10 calibrations ranged from 0.86-0.96, 23-53% of samples in these validations had predicted values that deviated 10% or more from their bioassay determined values. Subsequently, utilizing near infrared reflectance spectroscopy to create calibrations that were ingredient specific improved prediction accuracy for the 10 essential amino acids. For example, ingredient specific calibration curves for dried distillers’ grains improved R2 values to above 0.93 for all 10 essential amino acid validation curves and resulted in 16% or less of samples having predicted values deviating plus or minus 10% or greater from their bioassay determined values for all essential amino acids except lysine and tryptophan. These results indicate that near infrared reflectance spectroscopy has the potential to accurately predict the digestible amino acid content of feed ingredients, especially if ingredient specific calibration models are created for each amino acid.