Prediction of fatty acid composition of rectus abdmonis in beef using near infrared spectroscopy

National audience; The aim of the study was to test the prediction performance of the fatty acid (FA) composition of beef meat (Rectus abdominis, RA) with a NIRS (near infrared spectroscopy) laboratory equipment and to study the effect of different methods of sample preparation: without grinding (B0), ground with a meat chopper (B1), ground with a knife mill (B2) and finely ground in liquid nitrogen (B3). In order to have a representative database of French meat FA composition, 76 RA were sampled on carcasses from different rearing systems (dairy or beef, pasture or corn silage with or without extruded linseed concentrates. FA compositions were assessed using gas chromatography and the spectra were measured at wavelengths between 400 and 2500nm. The results show a better prediction of FA composition concomitant with an increased gradient of sample homogenization. NIRS accurately predicted several individual FA (16:0, 18:0, 18:1d9cis) and FA groups (total saturated and monounsaturated FA) (RPD>2.5). For other FA and especially polyunsaturated fatty acids (PUFA) n-6 or n-3, the performances were not satisfactory for quantitative purposes whatever the grinding method.