Novel applications of NIR spectroscopy for the optimization of monogastric animal diets

Identifier:  TDX:4305

Handle: https://hdl.handle.net/20.500.11797/TDX4305

Authors:  Cruz Conesa, Andrés

Abstract:
To maximize the productivity of monogastric animals (e.g. poultry and swine) and ensure balanced nutrition, it is crucial to consider the digestibility of nutrients in their feed. Bioassays are commonly used to determine how animals utilize feed, although they expensive, laborious, and generate a large number of samples (feed, faeces/excreta and/or intestinal contents) for analysis. Wet chemistry analytical methods are usually expensive and time-consuming. Near-infrared spectroscopy (NIRS) coupled with chemometrics offers a non-destructive, multi-parametric, and efficient alternative for nutrient analysis in the samples that bioassays involve. In this thesis NIRS was used to determine the nutrient content of various feedstuffs, including carbohydrates and lignin, as well as to predict accurately the nutrient composition of compound feeds, poultry excreta, pig faeces and poultry ileal digesta including amino acids. This will help reduce the cost of digestibility studies and improve the search for optimal feeds. To ensure a long term validation of the NIR models the new variability present in samples coming from new bioassays have to be accounted for. To address this, two strategies were proposed. The first strategy involves identifying interesting samples from compound feeds using diagnostic measures and periodically updating the calibration models. The second strategy selects a reduced number of new samples from a large batch of pig faeces using an efficient algorithm developed in-house, allowing accurate prediction of the remaining samples and incorporating new information into the model. Considering ethical concerns towards animal experimentation, the exploration of NIRS should extend beyond nutrient content determination, potentially replacing bioassays. In this regard, this thesis examined different strategies to predict nutrient and energy digestibility using NIRS. While varying degrees of success were achieved, these studies pave the way for future research in this area.