Rapid Determination Of Fatty Acids Nutritive Index In Infant Formula Powdered Milk Uising Near Infraed Spectroscopy

Fatty acids are generally divided into three broad categories, namely is saturatedfatty acid (SFA), monounsaturated fatty acid (MUFA) and polyunsaturated fatty acid(PUFA). Numerous studies have demonstrated that the category and proportion offatty acids in food are closely related to the health of human body nutritionalrequirements. Now, the well recognized recommended ratio value is that fatty acidsfor energy should not exceed than thirty percent of total energy in the world.According to it, the content ratio of SFA, MUFA and PUFA is1:1:1. In infantformula powdered milk, the high content of SFA is difficult to be absorbed and caninhibit the absorption of calcium; Monounsaturated fatty acids can provide energy,constitute the organizational structure, help nerve cells to repair, sustain the health ofthe brain and improve memory; Polyunsaturated fatty acids can enhance the growthand development of nerve and visual system in infants and young children, and alsopromote infant brain cell growth and mental development to improve memory. Lauricacid and Myristic acid are two kinds of saturated fatty acid in the body. They haveantibacterial and antiviral ability. Lauric acid is accounted for about5.8%in thebreast milk, and nutmeg acid in the mother’s milk is generally accounted for8.6oftotal cream. The provision in GB10765-2010National food safety standard Infantformula indicates that the total amount of lauric acid and myristic acid is less thantwenty percent of total fatty acids. At present, the commonly method used to detectthe contents of fatty acids in infant powdered milk is gas chromatography, etc. Thesemethods are time consuming, complex and costly. Compared with the conventionalmethods, Near Infrared spectroscopy (NIR) has many advantages, such as rapid,efficient and environmental protection. In this study, different brands infant formulapowdered milk was used. GC method determines the content of fatty acids in infantformula powdered milk as standard method. Spectrum of the samples is scanned usingSpectrum400FT-IR. The best mathematic model was established by selecting spectralrange and using different pretreatment. Finally the optimum model of the differentfatty acids nutrition indicators was determined, and while evaluating, the main resultsobtained are as follows:1. Acetyl chloride methanol methyl esterification method was used to processsamples. Gas chromatography was used to determinate content of each fatty acids in infant formula powdered milk. Acetyl chloride methanol methyl esterification methodwas used as reference test method, which determinates the chemical values of all thesamples to develop fatty acids model using the near infrared spectral. It provided datasupport for near infrared spectral rapid evaluation of the fatty acid composition ininfant formula powdered milk.2. PCR(Pdneipal Component Analysis) and PLS(Partial Least Square) predictionmodels were used for the ratio of the total amount of lauric acid and myristic acid andtotal fatty acids in infant powdered milk. Reliability of the models were verified bycross-validation and external-validation. From the results of PCR calibration model,the performance of the model had no obvious increase by choosing differentwavelengths and using different pretreatments. This showed that models might beconsidered to be insuffcient for most applications using NIRS. From the results ofPLS calibration model, the optimum NIR calibration model was obtained by selecting8554~4000cm-1region and SNV pretreatment. The best result was that coefficient ofdetermination in the model was0.9273and coefficient residual predictive deviation(RPD) was3.33. The established model was used to forecast the prediction set, bypaired sample test between the chemical methods and NIRS method in prediction set,P=0.114>0.05, which showed that there was no distinct statistical difference betweenthe two methods. The results showed that the method can be used to analyze the ratioof the total amount of lauric acid and myristic acid and total fatty acids in infantpowdered milk.3. Partial least squares regression prediction model for TFA, SFA, MFA andPUFA content of the infant powdered milk using near-infrared spectroscopy had beenestablished. Reliability of the model was verified by cross-validation and external-validation. Different wavelengths were selected and different pretreatments were usedto develop calibration models. From the results of PLS calibration model, theoptimum NIR calibration model was obtained by selecting8554~4000cm-1regionand SNV pretreatment. The best result was that coefficient of determination in themodel for TFA, SFA, MFA and PUFA content were0.9337,0.9374,0.9073and0.9203, respectively. And coefficient residual predictive deviation (RPD) were3.63,3.65,3.00and3.17, respectively. This proved that these models have a goodpredicting performance. Chemical and estimated values of correlation coefficient forTFA、 SFA、MUFA、PUFA were as follows:0.9847,0.8842,0.9474and0.9120inthe prediction results. It showed that there was no distinct statistical difference between the two methods by prepared sample test between the chemical methods andNIRS methods in prediction set. This proved that NIRS method was veracity to beused in practice to detect fatty acid content in infant powdered milk.The results showed that NIR calibration model is an effective method to predictthe different fatty acids nutrition indicators in infant powdered milk.