alpha-NH2 determination from biological infrared spectra

The use of Mid-infrared spectroscopy for the analysis of food products has only recently developped with the advent of Fourier Transformed spectroscopy and other techniques like Attenuated Total Reflectance, diffused reflection combined to the use of powerful micro-computers, We have recently reported to use a combination of multidimensional statistical analysis and Mid-FTIR spectroscopy for the quantitative determination of sugars in a biological sample. In this paper we have evaluated the use of such a method for the quantitative measurement of alpha-amino groups from amino-acids, peptides and proteins. The spectral region where the characteristic absorption bands of such groups are located, ranges between 1200 cm(-1) and 1900 cm-1. Water features a major absorption band in this region (1500-1700 cm-l). This superimpose with amide I and II bands. The standard deviation for each and every wavelength, calculated for all the spectra of the calibration set, show the existence of two absorption bands in the 1500-1900 cm-l region which means that the observed variations in this zone are not only due to water but are also due to two peaks centered at 1650 cm-1 and 1540 cm-1 (with a hollow at 1600 cm-1) that are characteristic of protids. The contribution of the first four axes of the PCA, axes 5, 1, 4 and 2, to the total inertia percentage are 2.37%, 53.36%, 3.92% and 28.82% respectively. The correlation coefficient between the major axis, axis 5, and the chemical values of alpha-NH2 is 0.311 and the second axis, axis 1, increases this value to 0.541. The first 10 axes were used to establish the prediction equation; the correlation coefficient value is very high : 0.978. Good predictions were obtained; mean and standard deviation associated to the predicted concentrations of alpha-NH2 content, valued 0 g/ml and 0.12 g/ml respectively. Hence, we have established the possibility of determining, from a MIR spectra, the alpha-NH2 content.