Predição do conteúdo de aminoácidos essenciais do grão de milho

Abstract

An experiment was conducted to evaluate the relationship between amino acids and nitrogen fractions of corn grain (Zea mays L.) and verify whether it is possible to use such relations to estimate amino acid contents. Forty five samples of corn hybrids cultivated in West of the State of Paraná were obtained and grouped in three categories according to the level of crude protein (low, medium and high). Afterwards, a random sub-sample was taken from each of the previous categories, resulting in a sub-sample of 15 corn hybrids that were taken out and submitted to amino acid and non-protein nitrogen (NNP) analysis. The content of corrected crude protein (CCP) was obtained by reducing the NNP concentration of CP values. The relation between variables was studied by Pearson Correlation and lineal regression models adjusted by the minimum square method. The angular coefficients of the first limiting amino acids for birds and swine (lysine, methionine and threonine) were positive and significant (LYS=0.1080+0.0140CP, R2=0.39; MET=0.1015+0.0084CP, R2=0.43, THR=0.1268+0.0184CP, R2=0.51). However, the determination coefficients of equations were lower than 0.51, showing that CP explains only part of the variation in the amino acid content of corn grain. The prediction of limiting amino acids did not improve when CCP was used as an independent variable. Based on the regression analysis, it is not possible to estimate precisely the content of essential amino acids of corn grain using as independent variables crude protein or corrected crude protein.