Deliverable D1.6 Predictive equations linking feed composition to the nutritive value to be applied with NIR technology

Objectives

The objective of task 1.6 was to evaluate the suitability of near-infrared spectroscopy (NIRS) for the determination of the nutritive value of individual components and mixed diets for pigs. The project utilises samples from a database collected over more than 40 years of animal experimentation to develop NIRS calibrations to predict macronutrients, amino acids, and the digestibility of nutrients and energy and metabolisable energy. The deliverable is based on activities in WP1 task 1.6.

Rationale

The nutritive value of a feed varies between and within feeds due to factors like genetics, agronomics, harvest, storage, and processing. In animal nutrition, adjustment of variation in the nutritive value is commonly done by analysing the nutrient fractions and using this information together with table values for their digestibility. However, it is known that the digestibility of nutrients may vary considerably from one feed to another and from one sample to another resulting in rather inaccurate measures of the nutritive value of the actual batches. Since it is not possible to perform in vivo evaluations of the nutritive value of individual batches because of time and cost constraints, there is a need for quick and reliable methods to determine the nutritive value of single feedstuffs for use in feed formulation and for control of complete feeds. Near-infrared spectroscopy (NIRS) has been widely used to evaluate the nutritional quality of agricultural products for several decades. While there are many examples on the use of NIRS to predict the chemical composition, less is known about its applicability to predict the nutritive value. In task 1.6, nutritionally evaluated samples collected for more than 40 years have been used to develop calibration models for macronutrients (i.e., ash, fat, protein, available carbohydrates, starch crude fibre, acid detergent fibre, and neutral detergent fibre) amino acids composition, and digestibility of energy and nutrient fractions and metabolisable energy. To have datasets with sufficient samples, feed types were grouped into ‘like’ sample types. All samples were also examined as one group (total) to determine if it was possible to have one calibration for all pig feeds.