| Five experiments were conducted to estimate the metabolizable energy(ME) of corn, corn DDGS and cassava with enzymatic hydrolyzate gross energy in roosters. The experiment 1 was conducted to investigate the effect of variation of endogenous energy loss(EEL) determined in different batches and seasons on the true metabolizable energy(TME) values of corn, corn DDGS and cassava. The experiments 2, 3 and 4 were to develop the linear models for estimating the ME of corn, corn DDGS and cassava with EHGE, respectively. The experiment 5 was to validate the additivity of the prediction models established in the experiments 2-4 by comparing the metabolizable energy values determined by force-feeding method and estimated by the models of corn, corn DDGS and cassava in the diets.Experiment 1 was to evaluate the effect of variation of EEL on the TME of feed ingredient to ensure the reproducibility and comparability of calculated TME of feedstuffs determined in different batches and seasons. A single factorial completely randomized design was adapted. One hundred and eight adult Hy-Line roosters were divided into 9 groups with 4 replicates of 3 birds in each. A total of 12 batches of bioassay were determined across the spring, summer and autumn. In each batch, two groups of rooster were selected to determine the EEL and the apparent metabolizable energy(AME) of 1 sample of 3 different sources of corn, corn DDGS, cassava slice and cassava meal, respectively. The mean of EEL determined in each season were calculated on the data from all batches of determination in the season. The values of true metabolizable energy of these feedstuffs were calculated from the AME of the corresponding feedstuffs and the mean of EEL which determined in experiment 1 in different seasons. The results showed as follows: 1) Significant differences were observed on EEL in all 12 batches(P<0.05), but no significant differences were observed on that determined in 3 seasons(P>0.05).Thus, the values of EEL from the batches within the same season can be merged for a mean of EEL in the season. 2) The EEL in autumn(16.25 Mcal/48 h per bird) was significant lower than spring(19.85 Mcal/48 h per bird) and summer(19.10 Mcal /48 h per bird)(P< 0.01). However, there is no difference in the EEL determined in spring and summer(P>0.05). 3) The EEL was highly positive correlated to dry endogenous feces and urine output in the same seasons(r≥0.91;P< 0.01). 4) The difference of the values of TME calculated with the EEL in different seasons for the same feed were ranged from 66.92 to 86.04 Kcal/g. In conclusion, there is a small variation of EEL determined among seasons, however, this variation had no significant effect on the true metabolizable energy of feed for roosters.Experiment 2 was conducted to develop a linear model to predict ME with EHGE of 10 corn samples to provide a rapid evaluation method for corn in practice. A single factorial completely randomized design was adapted. One hundred and eight adult Hy-Line roosters were divided into 9 groups with 4 replicates of 3 birds per replication. EHGE was determined with 5 replicates of 1 digestion tube per replication by a computer-controlling simulated digestive system developed by Institute of Animal Sciences, Chinese Academy of Agricultural Sciences. The ME was determined by force-feeding method at the same time. The result showed as follows: 1) the EHGE of 10 corn samples ranged from 3683 to 3877 Kcal/Kg with an average value of 3786±62 Kcal/Kg, the AME and TME ranged from 3392 to 3593 Kcal/Kg and 3707 to 4040 Kcal/Kg, respectively, the means were 3522±72 Kcal/Kg and 3932±94 Kcal/Kg, respectively. 2) The correlation coefficient between EHGE and AME and between EHGE and TME were 0.99 and 0.84, respectively. 3) The linear model of 10 corn samples for estimating AME from EHGE was AME=1.1569×EHGE-859(R2=0.9832;RSD=10 Kcal/Kg),and the regression equation for TME was TME=1.2772×EHGE-904(R2=0.7104;RSD=54 Kcal/Kg). 4) The RSD of the prediciton models for AME and TME were less than 17 Kcal/Kg and 93 Kcal/Kg, respectivily. 5) Compared with the models to predict ME from chemical composition, the prediction models for ME from EHGE determined with bionic method has a higher R2 and lower RSD.Experiment 3 was conducted to establish the linear model to estimate ME from EHGE of 9 corn DDGS samples in roosters. The method was in accordance with that in experiment 2. The result showed as follows: 1) the EHGE of 9 corn DDGS samples were ranged from 2921 to 3410 Kcal/Kg with a mean of 3164 ?186 Kcal/Kg. AME values were ranged from 2405 to 2929 Kcal/Kg with a mean of 2656? 183 Kcal/Kg. TME values were ranged from 2920 to 3390 Kcal/Kg with a mean of 3086?176 Kcal/Kg. 2) The EHGE was highly positive correlated to AME and TME(r>0.96). 3) The linear model for estimating ME of 9 corn DDGS samples from EHGE were AME=0.9611 × EHGE-385(R2=0.9464;RSD=40 Kcal/Kg),TME=0.8797×EHGE+302(R2=0.8519;RSD=70 Kcal/Kg). 4) The RSD of the prediciton models for AME and TME were less than 80 Kcal/Kg and 95 Kcal/Kg, respectively. 5) Compared with the models to predict ME from chemical composition, the prediction models for ME established from EHGE determined with bionic method has a higher R2 and lower RSD.Experiment 4 was to establish a linear model to evaluate ME from EHGE of 7 cassava samples including 3 cassava chips and 4 cassava meal in roosters. The method was in accordance with that in experiment 2. The result showed as follows: 1) The EHGE of 7 cassava was 2147 to 3479 Kcal/Kg, AME was 2170 to 3451 Kcal/Kg, TME was 2673 to 3905 Kcal/Kg. The EHGE, AME and TME of 3 cassava chips were significantly higher than 4 cassava meals(P < 0.05). 2) The EHGE was highly positive correlated to AME and TME(r>0.96), respectively. 3) The liner model for estimating ME of 7 cassava were AME=1.01×EHGE-58(R2=0.9886; RSD=71 Kcal/Kg) and TME=0.969×EHGE+500(R2=0.9830; RSD=84 Kcal/Kg); 4) The RSD of the prediciton models for AME and TME were less than 124 Kcal/Kg, respectively. 5) The similar accuracies for the models to predict ME from chemical composition and from EHGE determined with bionic method with the R2 up to 0.98 and the RSD less than 84 Kcal/Kg for each.The experiment 5 was to validate the additivity of the regression models developed in experiment 2, 3 and 4 by comparing the predicted and determined values of ME of 8 diets. The 8 diets consist of corn, corn DDGS and cassava at different concentration. The values of EHGE and ME of 8 diets were determined with the same methods described in the experiment 2. The calculated values of EHGE and ME of 8 diets were obtained according to the determined EHGE and ME of individual ingredients and their concentration in diets, respectively. The predicted values of ME of 8 diets were obtained according to the individual predicted values calculated from the EHGE of the ingredient and their concentration in diets. Linear regression was applied to analyze the fitness of estimated values and the determined values of ME of 8 diets to the line of Y=X. The result showed as follows: 1) Difference between the calculated and determined values for EHGE of the diets is not significant(P>0.05), which means additive. 2) Difference between the calculated and determined values for ME of the diets is also not significant(P>0.05), which means additive too; 3) There is a significant difference(P<0.05) between the estimated and determined values for AME of the diets, but no difference was observed on TME(P>0.05). The slope and the intercept of the linear model between estimated and determined ME showed no significant difference from 1 and 0(P>0.05), respectively. Thus, the model was coincident with the line of Y=X, which means that the additivity of the models was good for estimating the ME from EHGE of corn, corn DDGS and cassava.In summary, as a single factor, EHGE can be used to predict the ME values of feedstuffs. Compared with the chemical composition method, EHGE has a better correlation and a higher precision. |
PDF Full Text Request