Predição da energia metabolizável do milho para suínos com o uso da espectroscopia de refletância no infravemelho próximo

• Autor(a) principal: Ferreira, Silvia Letícia
• Data de Publicação: 2016
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da Universidade Estadual de Maringá (RI-UEM)
• Texto Completo: http://repositorio.uem.br:8080/jspui/handle/1/1796
• topic: Suínos; Nutrição; Milho; Energia metabolizável; Equação de predição; Espectroscopia de refletância no infravermelho (NIRS); Brasil.; Chemical composition; Corn; Metabolizable energy; NIRS,; Pigs; Brazil.; Ciências Agrárias; Zootecnia

The aim of this study was to determine the chemical composition and energetic values of different varieties of corn and evaluate mathematical models to predict the corn metabolizable energy values (ME) for pigs by using the Near Infrared Spectroscopy - NIRS. This study was performed in two stages: Stage 1: calibration curves were built to predict the chemical composition of the corn by using the NIRS technology. A total of 89 samples of corn were analyzed by wet chemical methods, and were also scanned by NIRS, in which 80 samples were used to establish the calibration model, and 9 samples were used for the external validation. The reading of the samples was carried out in the spectrum range between 1,100 and 2,500 nm and applied the regression model, and the model parameters estimated by the method of Modified Partial Least Squares (MPLS). For dry matter (DM), mineral matter (MM), neutral detergent fiber (NDF), acid detergent fiber (ADF), ether extract (EE), crude protein (CP) and gross energy (GE), the determination coefficients on the calibration set were 0.99; 0.30; 0.15; 0.40; 0.83; 0.94; 0.86 and the standard error of prediction (SEP) for them were 0.33; 0.16; 0.81; 0.43; 0.21; 0.21 e 14.06, respectively. The best residual performance deviation (RPD) were obtained for dry matter (RPD = 6.93) and for fat (RPD = 3.23). For the validation set the laboratorial values for ash ranged between 0.88% and 1.19%, and the values obtained in the NIRS ranged between 1.14 and 1.27%. For NDF the reference method indicated values between 10.91 and 13.39%, while the technology NIRS showed values between 11.77 and 12.35%. The ADF values ranged between 2.99 and 4.00% for the reference method and between 3.71 and 4.07% for the NIRS. For CP the reference values were between 7.71 and 9.91% and 8.08 and 8.81% for the NIRS. The reference values for fat and GE were between 3.12% and 4.24% and 4434 Kcal and 4602 Kcal respectively, and the NIRS indicated values between 3.49% and 4.07% and 4444 kcal/Kg and 4483 kcal/Kg, in this same order. The NIRS demonstrated to be effective to determinate the content of NDF, CP, GE and fat of corn being similar to the reference methods. Stage 2: a digestibility trial was carried out to determine the nutritional value of different corn cultivar. Forty-four crossbred barrows were used, distributed in a randomized blocks design with ten treatments plus the reference diet, and four replicates per treatment. The corn cultivars replaced 25% of the reference diet. Linear regression models of first degree of the estimated values of ME regarding the observed ME values were adjusted. The existence of linear relation was evaluated by detecting the significance to a posteriori estimates of the straight line parameters. The levels of DM, MM, CP, EE, NDF, ADF and GE for each corn cultivars was determined following laboratory reference methods, wich ranged from 86.57 to 89.71; 1.00 to 1.52; 8.28 to 10.71; 3.74 to 4.10; 11.08 to 13.83; 3.47 to 4.82; 4,390 to 4,478 respectively, and using the NIRS technology ranged from 87.69 to 91.07; 1.21 to 1.30; 9.65 to 10.96; 3.56 to 4.35; 11.69 to 12.29; 3.50 to 3.79 and 4,473 to 4,518 respectively. The values of digestible energy (DE) and metabolizable energy (ME) ranged from 3,400 to 3,752 and 3,244 to 3,611kcal/kg, respectively, and the digestibility and metabolizability coefficients of gross energy ranged from 85.68 to 95.66 and 83.00 to 92.07%. Eleven prediction equations were used to estimate the ME values, these equations had been also inserted into the NIRS to predict the energetic values of the corn by using the Near Infrared spectroscopy. The prediction equations EM1 = 4334 8,1MM + 4,1EE 3,7FDN; EM2 = 4194 - 9,2MM +1,0PB + 4,1EE 3,5FDN; EM7 = 344,272 + 0,90886EB + 57,9377EE - 86,9320PB e EM8 = 16,13 9,5FDN + 16EE + 23PB x FDN - 138MM x FDN were the most adequate to predict the metabolizable energy values of corn using NIRS