Defining feed wheat quality for broilers

Abstract

The aim of this project was to identify chemical and/or physical parameters which consistently related to the nutritive value of wheat in diets for poultry. To achieve this, selected physical and chemical parameters were measured in a wide range of wheat samples, and these were correlated with measurements of animal performance.

The possibility of using near infrared spectroscopy (NIRS) was then investigated as an effective method of screening the nutritive value of wheat.

This project was an expansion on a previous HGCA-funded project (Project Report No. 260). Chemical and physical parameters measured included specific weight, thousand grain weight, in vitro viscosity, gross energy, nitrogen, neutral detergent fibre (NDF), starch, total and soluble non-starch polysaccharides (NSP), lysine, threonine, amylose, grain hardness, rate of starch digestion and protein profiles.

Animal performance parameters measured included dry matter intake (DMI), live weight gain (LWG), gain:feed, in vivo viscosity, apparent metabolisable energy (AME), ME:gain, dry matter retention, ileal dry matter digestibility, ileal starch digestibility and ileal protein digestibility.

The wheat database used in this study had wide-ranging chemical and physical parameters, leading to all bird performance parameters being significantly affected by wheat sample. Fungicide and fertiliser application and seed rate affected wheat nutritive value. When correlations between wheat parameters and bird performance were calculated, specific weight was not significantly related to performance (r = 0.031, 0.008 and -0.019 for DMI, LWG and gain:feed respectively). However, thousand grain weight, hardness, the rate of starch digestion, and in vitro viscosity were weakly related (r < 0.523 for DMI, LWG and gain:feed).

Using NIRS, the chemical parameters of wheat are best predicted by scanning milled wheat, while the nutritive value of wheat was best predicted by scanning the undried whole kernel wheat samples. Calibrations for milled samples produced acceptable (>0.75) coefficients of cross validation (R2cv) for specific weight, crude protein and rate of starch digestion, while calibrations for whole wheat samples, undried, produced acceptable R2cv for total live weight gain and gain:feed.

In conclusion, the choice of wheat fed to broiler chickens substantially affected bird performance. It appears that NIRS has the potential to be an effective and rapid technique for establishing wheat nutritive value; additionally the wheat can be scanned on an "as is" basis for more rapid determination. A larger database of samples would be recommended to fully test this concept.