Alternative methods to predict the nutritive value of broiler chicken diets : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Animal Science at Massey University, Palmerston North, Manawatū, New Zealand

Abstract

Feed represents the major cost of poultry production accounting for up to 70% of the total. Accurate estimation of the chemical composition and nutritive value of the feed ingredients and diets is essential to reduce costs and optimise diets. Undersupply of nutrients will have a strong negative impact on the performance of birds and oversupply will have a negative impact on the environment as well as cost of production. Traditional methods of feed evaluation are time-consuming and costly. Both in vivo and in vitro techniques have several limitations. Therefore, there is a need for rapid and accurate analysis of feed samples for their chemical composition and nutritive value in real-time for optimal formulation of diets. Prediction equations and Attenuated Total Reflectance Fourier Transform Mid-Infrared Spectroscopy (ATR-FT MIRS), however, could offer a much faster approach to predict the chemical composition and nutritive value of broiler diets. Therefore, this thesis aimed to develop and validate prediction equations to be applied in practical conditions and to investigate the potential of ATR-FT MIRS in predicting chemical composition and nutritive value of broiler diets. The first study (Chapter 3) was designed to develop prediction equations using stepwise regressions. Bootstrap was used to select the variables as well as to confirm the stability of the models. The results indicated that prediction models for coefficient of apparent ileal digestibility (CAID) and ileal digestible content (IDC) of nutrients could be developed using the stepwise regression and bootstrapping approach. To ensure the practical application of each equation with greater confidence and to alert users to potential risks, Chapter 4 was designed to validate the proposed prediction models developed in Chapter 3. The validation revealed that some equations could not be used in all situations (CAID of nitrogen (N), fat, starch, and calcium (Ca) and IDC of Ca), and some could be used as a general guide (CAID and IDC of phosphorus (P)). The equations developed for CAID of energy and dry matter (DM), and IDC of N, starch, energy, and DM could be used with good results in terms of R2 and CCC in the validation study. The 5th, 6th, and 7th Chapters investigated the potential use of ATR-FT MIRS in predicting the chemical composition and nutritive value of broiler diets using various approaches. Many of the wet chemical compounds in broiler diets, ileal digesta, and excreta were predicted well. The results also revealed that the nutritive value of broiler diets can be predicted from the diets itself, ileal digesta, and excreta spectra as well as the combination of diet and ileal digesta spectra or diet and excreta spectra. Overall, the results of this thesis indicated that prediction equations and ATR-FT MIRS can be used as a rapid real-time technique to evaluate the nutritive value of broiler diets. The findings of this thesis regarding ATR- FT MIRS are novel and can be used for further update and development of MIRS calibration models, which could substantially reduce the cost of wet chemical analysis in future.