The development of digestibility coefficient database of rapeseed meal

Abstract

 The aim of the study was to establish the variation in chemical composition and amino acid concentration of rapeseed meal (RSM) and develop near infrared reflectance spectroscopy (NIRS) equations to predict ileal and total tract digestibility of RSM for pigs and poultry. Ninety two samples of RSM (representative of the range in crude protein (CP) and neutral detergent fibre (NDF) content of commercially available RSM) were collected over a two-year period, scanned by two NIRS instruments and transferred to AFBI for wet chemistry analysis (CP, NDF and amino acids) and for formulation into pig and broiler diets. NIRS 1 was based at Aunir and was a NIRSystems 6500 spectrophotometer (Foss). NIRS 2 was based at QUB and was an Antaris II FT-NIR instrument (Thermo Fisher Scientific). Pig and broiler diets contained 500g/kg RSM as the only protein source and energy was supplied in the form of maize starch and dextrose. Nitrogen-free diets were formulated to enable measurement of basal endogenous losses of amino acids. For pigs, the diets were offered to eight batches of 12 post value T caecum cannulated pigs (Landrace x Large White) in a partially balanced, eight period changeover design. For broilers, diets were offered to 15 batches of 64 male broilers (Ross 308) over 15 experimental trials.  Each batch contained two controls (the N-free diet and RSM3) in order to compare results between batches. Therefore, there were eight diets offered to eight birds in each batch. Apparent ileal digestibility (AID) of dry matter (DM) and amino acids for pigs and broilers was determined as was standardised ileal digestibility (SID) of amino acids using the basal endogenous losses to calculate. Total tract digestibility of DM, energy and NDF were also determined for pigs and broilers.

There was wide variation in CP, NDF and amino acid content of RSM available for use in animal feed in the UK indicating the need for a means to accurately predict the nutritive value of RSM when offered to pigs and broilers. NIRS was found to accurately predict CP, NDF and lysine content of RSM and has the potential to predict the content of other amino acids. Digestibility co-efficients (AID, SID and total tract) were also highly variable with variation in some parameters over 70%, again highlighting the requirement for an alternative to average “book” values within formulation packages. A large database of digestibility values (AID, SID and total tract) for RSM has been established. NIRS has the potential to predict AID and SID of some amino acids (particularly lysine in pigs and threonine in broilers) with ratio to prediction (RPD) values being greater than >1.5. However, it must be noted that the RPD values were generally <1.5 and, as such, suggest that the NIRS prediction calibration is unsatisfactory. Future work should focus on those particular amino acids where NIRS prediction of digestibility resulted in RPD values of greater than 1.5. These NIRS prediction equations could be further developed by increasing the dataset. Measured basal endogenous losses were successfully determined and were in line with published values which will provide a useful basis for future research on different feed ingredients.