Near infrared reflectance analysis of grains to estimate nutritional value for chickens

Maximising bird growth rate, feed efficiency and profitability of broiler production depends, among other factors, on accurate measurements of the apparent metabolisable energy (AME) content (MJ/kg) and AME intake (MJ/d) of cereal grain based diets. Near infrared (NIR) calibrations have been developed to measure these characteristics for cereal grains. Further research is being conducted to strengthen the calibrations so they can become the primary basis for trading grains for livestock in Australia.

Apparent metabolisable energy values for ingredients are used by the broiler industry to formulate diets that meet predetermined energy density (MJ/kg) specifications. A single mean estimate for the AME value for each grain species is used commonly during diet formulation. However, research from the Premium Grains for Livestock Program (PGLP) showed that there was a large range in AME values (MJ/kg DM) for Australian sourced grains; being 11.9-15.3 for wheat, 10.9-13.6 for barley, 12.1-14.5 for triticale and 15.3-16.7 for sorghum (Black et al., 2005). Traditionally, glucanase and/or xylanase enzymes are added to broiler diets to reduce the range and increase the absolute values for AME in broiler diets. Recent analysis of the effects of addition of xylanase and phytase enzymes to diets formulated from 38 wheat, 8 triticale and 3 sorghum samples from PGLP showed that AME values were actually depressed by the enzymes for 12 wheat, 3 triticale and all 3 sorghum based diets (Black, 2008). Although the addition of enzymes raised the mean AME content of the wheat based diets by approximately 0.5 MJ/kg, the variation in AME values across the diets remained almost constant at 4.13 MJ/kg for diets without enzymes and 4.22 MJ/kg for diets with enzymes.
Several other important results for the broiler industry were obtained from PGLP (Black, 2008). First, there was no relationship between the AME content of a diet (MJ/kg) and the amount of the diet consumed by broiler chickens (R2 = 0.003). This result suggests that different characteristics of the grain determine digestibility compared with intake. Secondly, there were significant differences (P < 0.05) within grain types in the intake (g/d) by broilers when grain samples were incorporated into diets. For example, the intake of wheat based diets by broilers varied by 20% depending on the particular wheat sample incorporated at a constant proportion into the diet. The daily intake of AME (MJ/d) by broilers varied by approximately 34% across the wheat based diets. Thirdly, broiler growth rate was more closely related to AME intake (MJ/d) than to the AME content of the diet (MJ/kg). Cereal grains with their high starch content are the major energy source for broilers and represent from 60-70% of the diet. The results from PGLP and supported by the work from Scott (2005) indicate that values must be obtained for both the AME content (MJ/kg) and AME intake (MJ/d) to fully describe the energy value of any batch of cereal grain for broilers. Estimates of the money value of a 1 MJ/kg difference in the AME content of grain range from $11.50/t to $27/t depending on the base cost of the grain relative to other high and low energy ingredients (Black, 2008). Similarly, an increase in AME intake (MJ/d) that simulates growth rate and results in chickens reaching sale weight one day earlier has been estimated to be worth $2m/year for a 1 million bird per week broiler operation (Black, 2008). Rapid methods for measuring both the AME content of a grain and the relative AME intake of that grain compared with other grains when incorporated into a diet would be of great value to the broiler industry. In addition, a rapid method for estimating the effect of enzyme addition on the AME content and AME intake of grains would help identify when adding enzymes may be disadvantageous.

Near infrared spectroscopy (NIR) technology is now used widely to predict many chemical components of cereal grains. NIR has been applied to the results from PGLP to predict the AME content and AME intake index for any batch of grain for broiler chickens. The AME intake index was calculated by dividing the AME intake value (MJ/d) for every grain fed to broilers in PGLP by the highest value and multiplying by 100 to give values potentially from 1 to over 100. The AME intake index was used to provide a relative estimate of the likely intake for a broiler diet based on a particular grain rather than an absolute value in MJ/d which changes as chickens grow. The NIR calibrations were established across grain species and include all grains fed to broilers in PGLP. The NIR scans were on samples of whole grain rather than milled grain to reduce the cost and time taken for the analysis. The relationship between NIR predicted values and observed AME (MJ/kg as fed) for broilers is presented in Figure 1. The dashed lines represent ± 1 standard deviation (SD) from the observed mean values with individual grains predicted to be outside this range identified.

Figure 1 Relationship between observed and NIR predicted AME for grain based diets fed to broilers. 

1-VR (1-Variance Ratio) is the fraction of the variance in observations accounted for when some of the observations are used for ‘cross validation’ as determined by the calibration software. A value of 0.83 indicates acceptable robustness of the calibration. The value of the calibration for predicting unknown samples is assessed by (RPD) the Ratio of Prediction to experimental Deviation (SD/SECV) = 2.4. The calibration is rated as ‘quantitative’ with predictions being within ± 0.47 MJ/kg DM in 95% of samples measured. The relationship between NIR predicted and observed AME intake index values for broilers is presented in Figure 2. The value of the calibration as assessed by RPD was 1.8. The calibration is rated as ‘useful’ distinguishing between high and low values with predictions being within ± 4.96 index units in 95% of samples measured.

Figure 2 Relationship between observed and NIR predicted AME intake index values for grain based diets fed to broilers.

Figures 1 and 2 show that grain samples differing in energy value can be identified using the NIR technology both within and across grain species. A case study was conducted where samples of wheat from 37 locations in southern Australia were collected in 2005 and their energy value determined using NIR technology (Spragg, 2007). Two samples were selected, one with high energy (AME 13.1 MJ/kg as fed; AME intake index 69.3) and one with low energy (AME 12.5 MJ/kg as fed; AME intake index 64.2). These grains were compared, with and without enzymes, and with a wheat provided by a broiler company (AME 12.8 MJ/kg as fed; AME intake index 67.4) with enzymes in a full broiler growth study. The average number of days for male and female birds to reach 2.45 kg was 38.0, 37.8, 36.0, 35.8 and 37.1, respectively for the low energy wheat without and with enzymes, the high energy wheat without and with enzymes and the company wheat with enzymes. The study showed that NIR selected high-energy wheat reduced the time to sale by 2 days compared with the low energy wheat and by 1.3 days compared to the company selected wheat. These results suggest there is value in improving the accuracy and reliability of the calibrations for commercial use.

The Rural Industries R&D Corporation is currently funding two projects to expand the number and type of grain samples included in the calibrations. In one study, 90 cereal grains will be selected from samples provided from broiler companies and the Pork Cooperative Research Centre (CRC). These grains will include high screening samples, grains grown under irrigation, red wheat varieties, experimentally sprouted and water stressed grains. The second study will examine a range of new triticale cultivars. AME results and associated NIR scans of grains will be added to the existing PGLP database to improve the calibrations. V. DELIVERY OF CALIBRATIONS TO INDUSTRY The Pork CRC is also enhancing the NIR calibrations for measuring the energy value of cereal grains for pigs. The Grains R&D Corporation has licensed the Pork CRC to make the PGLP and enhanced calibrations for assessing the energy value of cereal grains for broilers, pigs and ruminants commercially available across the feed grain value chain. Sub-licences are being provided to the major feed testing laboratories, livestock integrator companies, stockfeed manufacturers and grain handling and broking companies across Australia. The NIR calibrations will provide a more suitable method for assessing the energy value of grain for livestock than the current methods based on test weight (kg/hl) and screenings percentage. The NIR calibrations are proposed to be used as the primary basis for trading grains for livestock in Australia.

Black JL (2008) Premium Grains for Livestock Program: Component 1 – Coordination. Final Report. Grains R&D Corporation, Canberra, Australia.
Black JL, Hughes RJ, Nielsen SG, Tredrea AM, MacAlpine R, van Barneveld RJ (2005) Proceedings, Australian Poultry Science Symposium 17, 21-29.
Scott TA (2004) Proceedings, Australian Poultry Science Symposium 16, 9-16.
Scott TA (2005) Recent Advances in Animal Nutrition in Australia 15, 237-244.
Spragg JC (2007) PGLP Technology Transfer and Commercialisation. Final Report. Grains R&D Corporation, Canberra, Australia