Use of near infrared transmittance spectroscopy to predict fatty acid composition of chicken meat

In the present study, near infrared transmittance (NIT) spectroscopy was tested as a potential analytical technique to predict the FA profile of ground chicken breast (Pectoralis superficialis) considering the wavelengths between 850 and 1050 nm. Calibration equations were built using reference data expressed as (i) percentage of total FA and (ii) absolute concentration, i.e., mg of FA in 100 g of fresh meat. Calibrations developed on FA expressed in absolute concentration were much more accurate than those developed on percentage of total FA. The highest performances were obtained for two groups of FA, namely saturated and monounsaturated (R-CV(2) of 0.90 and 0.93, and SECV, of 53.59 and 70.13 mg of FA on 100 g of meat, respectively), and for few individual FA (palmitic acid, oleic acid), having R-CV(2) higher than 0.9. Sample pre-processing (milling) and different spectra pre-treatments were necessary to maximise the performance. Polyunsaturated FA were the hardest components to determine (R-CV(2) of 0.62 and SECV of 61.96 mg of FA on 100 g of meat) and results suggest that NIT spectroscopy cannot be referred to as reliable method to predict these constituents. (C) 2012 Elsevier Ltd. All rights reserved.