Predicting conjugated linoleic acid (CLA) composition in Brazilian dairy products by multiple regression analysis based models

Abstract

The analysis of conjugated linoleic acids (CLA) in dairy products is of concern for human health, but requires a combination of different analytical methods that are laborious and time consuming. The aims of the present study were to develop mathematical models based on data of the CLA and fatty acid composition of Brazilian dairy products and to apply the developed models for the calculation of CLA based on a basic analysis of fatty acid composition. The fatty acid composition of five types of dairy products was determined by gas chromatography (GC); analysis of CLA was achieved by silver-ion high-performance liquid chromatography and GC-mass spectrometry. Stepwise multiple regression analysis was performed to model the composition of CLA. All the six significant models of CLA composition (total CLA; total cis,trans, trans,cis CLA; cis-9,trans-11 CLA; trans-7,cis-9 CLA; trans-11,cis-13 CLA; and total trans,trans CLA) were physiologically consistent, had appropriate goodness of fit, and were fairly robust. According to the results of Pearson correlation analysis between estimates and experimental data (r = 0.78–0.91 and 0.0000 < P < 0.0028), the models could be used to rank dairy products according to the estimates of total CLA, trans-7,cis-9 CLA; trans-11,cis-13 CLA; total cis,trans, trans,cis CLA; and total trans,trans CLA, but not for cis-9,trans-11. The adjusted regression models allowed the estimation of CLA composition in a variety of Brazilian dairy products.