Considering the milk fatty acid influence on human health, the aim of this study was to compare gas chromatography
(GC) and Fourier transform infrared (FTIR) spectroscopy for the determination of these compounds.
Fatty acid content (g/100 g of fat) were obtained by both methods and compared through
Pearson’s correlation, linear Bayesian regression, and the Bland–Altman method. Despite the high correlations
between the measurements (r = 0.60–0.92), the regression coefficient values indicated higher
measures for palmitic acid, oleic acid, unsaturated and monounsaturated fatty acids and lower values
for stearic acid, saturated and polyunsaturated fatty acids estimated by GC in comparison to FTIR results.
This inequality was confirmed in the Bland–Altman test, with an average bias varying from 8.65 to
6.91 g/100 g of fat. However, the inclusion of 94% of the samples into the concordance limits suggested
that the variability of the differences between the methods was constant throughout the range of measurement.
Therefore, despite the inequality between the estimates, the methods displayed the same pattern
of milk fat composition, allowing similar conclusions about the milk samples under evaluation.