Physicochemical Characteristics and Fatty Acid Composition

Traditional Tibetan Ghee: Physicochemical Characteristics and Fatty Acid Composition Bingyu Jing1, Wenjie Chen2, Mengzhu Wang1, Xiaohui Mao1, Jia Chen1, and Xiuzhu Yu1＊ 1 College of Food Science and Engineering, Northwest A and F University, 22 Xinong Road, Yangling 712100, Shaanxi, P. R. CHINA 2 Hybrid Rape Research Center of Shaanxi Province, 6 West Gaogangqu Road,Yangling 712100, Shaanxi, P. R. CHINA

Abstract: Fifty traditional Tibetan ghee (TTG) varieties were collected and analyzed their systematic characteristic indices, including physicochemical parameters, minerals, fatty acid composition, and thermal behavior. Results show that TTG contains a large amount of fat (71.68%–93.3%) and a small quantity of protein (0.51%–1.81%). The acid and peroxide values of TTG vary from 0.02 to 1.30 mg/g and 0.07 to 5.93 meq/kg, respectively. The content of minerals varied with altitude level and region significantly (p < 0.05), and the regional variations of fatty acids in TTG were also observed significantly, these differences may be due to the high unsaturated fatty acids level in the cow diets.

Key words: traditional Tibetan ghee, physicochemical characteristics, minerals, fatty acid composition

1 Introduction logical properties, it indicated that ghee produced from Ghee is a traditional Tibetan dairy product that is pro- cows’ milk based on grass-feeding have a potentially duced from a series of fermentation, heat clarification, and healthier fat content and a desirable flavour5）. Neupaney et desiccation of yak milk. The yak is placed under the sub- al. studied some functional properties of yak butter lipids family Bovine and belongs to the classification of Bos through analysis of conjugated linoleic acid（CLA）composi- grunniens, which is mainly found in the highlands of Nep- tion, differential scanning calorimetric（DSC）analysis, and alese Himalayas, Indian Kasmir, Tibet, Mongolia, and determination of tyrosine inhibition activity and antioxi- Bhutan1）. The major constituents of yak milk are fat（6％ dant property2）. Orrashid et al. studied the fatty acid –10％）, protein（5.5％）, lactose（5％）, and minerals（1.1％）2）. content of both domesticated and stocked cheese, indicat- Yak milk is increasingly recognized for its nutritional bene- ing that cheese that grazed on Himalayan alpine pastures fits and is included in products such as cheese, butter milk has a more balanced fatty acid composition3）, however, to tea, and yogurt. Due to the attractive appearance, grainy the best of our knowledge, the studies on characteristics of and semisolid texture, pleasant odor, excellent taste, and TTG were still lacking, while all of them are important pa- the high proportion of polyunsaturated fatty acids（PUFAs）3）, rameters for processing and utilization. The main objec- traditional Tibetan ghee（TTG）has important culinary pur- tives of this study are to analyze the global character such poses, such as dressing, cooking, frying, and even in reli- as physicochemical characteristics, fatty acid composition gious rites. Many food manufacturers and research organi- and mineral content, Fourier transform infrared（FTIR） zations also consider TTG as PUFA resources to meet the spectra, and thermal characteristics, most importantly, the increased consumer demand of PUFA-rich oils. relationship of geographical location with the content of Over the past several years, various studies have been minerals and fatty acid also studied in the paper. focused on the characteristics of TTG. Chassaing et al. in- vestigated the European ghee with industrial ghee in terms of peroxide value, minerals, and vitamin A, but the numbers of samples are insufficient to determine the 2 Materials and Methods physico-chemical indicators in ghee varieties4）. Moreover, 2.1 Samples many researches also focused on the functional properties The 50 ghee samples were collected from seven different in ghee. Silva et al. mainly studied the fatty acid composi- Tibetan regions, Sannan（n＝12, 3700 m）, Shigatse（n＝12, tion, physicochemical characteristics, as well as microbio- 4000 m）, Lhasa（n＝9, 3650 m）, Qamdo（n＝11, 3500 m）,

＊Correspondence to: Xiuzhu Yu, College of Food Science and Engineering, Northwest A and F University, 22 Xinong Road, Yangling 712100, Shaanxi, P. R. CHINA E-mail: [email protected] Accepted June 4, 2019 (received for review January 30, 2019) Journal of Oleo Science ISSN 1345-8957 print / ISSN 1347-3352 online http://www.jstage.jst.go.jp/browse/jos/ http://mc.manusriptcentral.com/jjocs

827 B. Jing, W. Chen, M. Wang et al.

Nagqu（n＝2, 3000 m）and Ngari（n＝4, 4500 m）respective- cooling to －30℃, held for another 10 min, followed by ly. All the yak milk were collected during mid-April period, heating cycle to 80℃ at the same rate. The cooling and the samples were stored and frozen（－18℃）before further heating cycle were repeated. The onset, peak, end temper- analysis. The specific information of samples is presented atures, and the enthalpy of thermal transition（melting and in Table A. 3. crystallization）were determined through the acquired thermograms. 2.2 Physicochemical analysis The acid value（AV）and peroxide value（PV）were deter- 2.5 Mineral analysis mined according to the standard methods of American Oil The analytical wavelengths（nm）chosen were: Al, Ca, Fe, Chemist’s Society（AOCS Official Method Cd 8b-90） Mg Na, Zn. For the operating conditions of the equipment, （AOCS, 2003a）. Briefly, the determination of PV mainly the radio frequency power（1.5 kW）and the plasma（15.0 L/ uses mixture of oil and chloroform/acetic acid, which was min）, auxiliary（0.4 L/min）, and nebulizer（0.7 L/min）gas left to react with a solution of potassium iodide indarkness; flow-rates were optimized from previous works. Single-ele- the free iodine then titrated with a sodium thiosulfate solu- ment standard stock solutions containing 1000 mg/L Cu, tion. The determination of AV mainly uses mixture of oil Fe, Mn, Mo, and Zn and 4000 mg/L Ca, K, Mg, Na, and P and diethyl ether/ethanol added with 2 mL phenolphtha- were suitably diluted to prepare analytical calibration solu- lein. The mixture was titrated with 0.01 M of potassium hy- tions. Elemental determination by ICP OES was performed droxide and shaken vigorously until the color was changed using external calibration. The standard solutions for the （from white to pink）. calibration curves were prepared at the same acid concen- Moisture was determined according to AOCS Official tration for each digestion procedure. Method 926.12, briefly, moisture content was determined from the loss of mass of a specified quantity of oil heated at 2.6 Fatty acid composition analysis 105±1℃ in a hot-air oven for 1 h, and crude fat and The fatty acid composition of TTG was determined protein were determined according to the standard method through conventional saponification/methylation proce- of the Association of Official Analysis Chemists（AOAC Of- dure. First, 60 mg of sample was mixed with 200 μL of 2％ ficial Methods of Analysis, 15th, 1990）. Protein content was methanolic KOH in screw-capped glass tubes. The tubes calculated from nitrogen content as follows: N（％）×6.25 were stirred for 5 min and held at room temperature for 30

using Kjeldahl’s method. The fat content was measured by s. After acidification with 1 g of NaHSO4, the released fatty using a Soxhelt extraction apparatus with petroleum ether acids were extracted twice by using 4 mL of isooctane.

as a solvent for 8 h. All analyses were carried out in tripli- Then, 2 mL of 5 mg/mL internal standard（C11: 0）was added cate. into the fatty acid mixture and vortexed for 1 min at ambient temperature. The fatty acid methyl esters 2.3 FTIR spectra analysis （FAMEs）were then extracted by the addition of 4mL of A Bruker VERTEX 70 series FTIR spectrometer（Bruker isooctane. This isooctane solution containing FAMEs was Optics, Germany）equipped with a deuterated detector was subjected to Gas chromatography–Flame Ionization used to obtain the FTIR spectra. At the ambient tempera- Detector（GC–FID）analysis to quantify the fatty acids. ture, a droplet of the sample was evenly smeared on the Under the same detection conditions, the retention time of screen of 100 meshes and placed in the sample chamber- mono-fatty acid methyl ester was compared with the stan- maid with at 50℃. The spectra were obtained from 6000 dard to identify the quality of fatty acids and with the area cm－1 to 400 cm－1 at a resolution of 8 cm－1 by using 64 co- of each peak for the quantitative analysis. GC measure- added scans. The background air spectrum was subtracted ments were carried out on a fused-silica capillary column before analyzing the sample spectra. （HP-INNOWAX, 0.25 mm×0.25 μL×30 m）. The temperature of the inlet was set at 250℃. After holding the sample 2.4 Thermal characteristic analysis at ambient temperature for 2 min, the oven temperature The thermal characteristics of samples were determined was increased to 240℃ at 10℃ min－1 and maintained for with a modulated differential scanning calorimeter（DSC, 10 min. High-purity nitrogen was used as carrier gas at flow TAQ20, USA）. The instrument was calibrated using indium rate of 1.1 mL/min. One μL of sample was injected by the （156.6℃, 28.57 J/g）and zinc（419.5℃, 108 J/g）. An empty split mode injection system.

DSC-pan was used as a reference. N2 was used as the purge gas at a flow rate of 40 mL min－1. The sample（6–9 mg）was 2.7 Statistical analysis accurately weighed directly on the DSC-pan（SFI-Alumi- All experimental measurements were conducted at least num, TA Instrument T11024）. The sample was allowed to in triplicate, and the data are expressed as mean±stan- equilibrate at 80℃ for 10 min after being heated to 80℃ at dard deviation by using Microsoft excel 2010, Spss and a rate of 5℃ min－1. Then, it was held for 10 min before Minitab software.

828 J. Oleo Sci. 68, (9) 827-835 (2019) Characteristics of Traditional Tibetan Ghee

3 Results and Discussion 3.1 Physicochemical analysis As shown in Table 1, chemical analysis revealed that the quality parameters such as AV and PV of TTG were 0.22 mg/g oil and 1.31 meq/kg respectively, and the PV of TTG is in the range for traditional Tunisian butter（1.12-1.67 meq/kg）6）, which is slightly lower than that of Eastern Ana- tolian ghee（1.28-9.83 meq/kg）7）, and the traditional East Algeria butter（1.60-3.01 meq/kg）8）. The AV of TTG is lower than that of Qinghai ghee（1.23-5.51 mg/g）. The low value indicates that TTG possesses lower quantities of free fatty acids（FFAs）and oxidation by-products. Findik et al. indicated that any oil with PV ＞15 meq/kg oil is considered 8, 9） unfit for human consumption . Due to the high propor- Fig. 1 DSC thermograms of TTG. tion of highly oxidation-sensitive PUFAs, a continuous in- crease in PV was observed in TTG during storage. There- peak, which is mainly due to the presence of the mixture of fore, the protection of samples from possible oxidation triglycerides12）. During melting, the 3 samples show a degradation by incorporating antioxidants is essential to similar broad exothermic peak on the melting curve from extend the shelf-life. 10-20℃（ Sanan）, 15-23℃（ Lhasa）and 10-35℃（ Qamdo）. Fat is an important factor to determine the quality of oil. The broadest exothermic peak range was observed in The crude fat in TTG varies from 71.68-93.39 g/100g oil, Qamdo ghee, which is mostly attributed to the manage- which is slightly lower than the range of 90.11-99.01 g/100 ment system, climate, season, and sampling periods13）. The g oil as reported for various ghee10）, and the high propor- melting of triglyceride also depends on its chain length, tion of fat will lead rancidity and generate rancid flavor extent of branching of the chain, the unsaturation of its when stored in unsuitable condition. Furthermore, the constituent fatty acids as well as their stereo-specific dis- protein content was 1.06 g/100 g oil in the samples, which tribution along the glycerol molecules14）, and the melting is within the range of 0.10-1.41 g/100 g oil as reported for point of oil increases with increasing chain length and satu- various kinds of TTG7）. ration of constituent fatty acids15）. As shown in Fig. 1, a The moisture content in TTG is within 10.55-30.41 g/100 sharp and intense peak appears at －15℃, indicating that g oil, which is in the same range as East Algerian butter TTG is a mixture rather than a pure semisolid. （17.50-26.01 g/100 g）8）. High moisture in TTG can lead hy- drolytic breakdown and generate rancid flavor. 3.3 FTIR spectra analysis The characteristic peaks in FTIR spectra can specifically 3.2 Thermal characteristic analysis represent the common characteristic molecular features of DSC has been used to study the thermo-oxidative pro- oil16）. Figure 2 presents the representative FTIR spectra of cesses that occurred in the oil as a function of temperature TTG, which were collected from the region of Sanan, Shi- and heat flow11）. Figure 1 shows the thermal characteristics gatse, Lhasa, and Qamdo in Tibet. The spectra in the high- of TTG samples collected from the regions of Sannan, wavenumber region present strong multiple bands. The Lhasa, and Qamdo in Tibet. The crystallization peak was 2975–2830 cm－1 range is assigned to C–H stretching of the not prominent when cooled to －30℃, but when the oil methyl and methylene backbones of the lipids. The distinc- was heated, the melting curve showed a broad peak from tive peak at 3468 cm－1 shown in the Fig. 2, was caused by －30℃ to 80℃. Moreover, the endothermic peak shows a the C-H stretching of C＝O bonds. Commonly, this peak in broad range from 0-30℃ rather than a sharp and intense higher wave-number region can be used to determine oxi-

a Table 1 Physicochemical characteristics of TTG . Parameters Mean Coefficient of variation (%) Minimum Maximum AV (mg/g) 0.22 1.22 0.02 0.27 PV (meq /kg) 1.31 0.88 0.07 5.93 Fat (g/100 g) 82.54 0.06 71.68 93.39 Protein (g/100 g) 1.06 0.30 0.31 1.80 Moisture (g/100 g) 14.5 0.35 11.2 17.9 a mean±standard deviations.

829 J. Oleo Sci. 68, (9) 827-835 (2019) B. Jing, W. Chen, M. Wang et al.

the essential mineral of the human body, showed the significant regional variation（p＜0.05）, specially 26.89 mg/L in Ngari, 17.99 mg/L in Lhasa, the concentration of magnesium decreased as the altitude level lower, this is mainly due to the various plants in different altitude level, finally lead to the variation of minerals. Sodium is an important inorganic element in human body, and the highest concentration of sodium is detected in Ngari, this may mainly linked to the local characteristics of large amount of precipitation and numerous lakes. For the microelements, the contents of Fe, Zn show the significant regional variation（p ＜0.05）, while the variation in altitude level is insignificant. As studied, the concentration of minerals closely linked to the agrotype, moisture content and organic matter21）. Fig. 2 FTIR spectra of TTG. Overall, all the highest concentration of minerals were found in Ngari, the local soil environment and vegetation dation degree of fats and oils17）. The sample in Shigatse may explain the high minerals level of TTG, the region shows the most intense stretching, which indicated that it which over 4500 meters are covered with the brushwood, possesses the highest unsaturation, the regional difference prairie and shrub and the local soil are rich in the cherno- in unsaturation is mostly attributed to the regional weather, zem, alpine meadow soil, with the high content of organic pasture sites and grass species. Moreover, the high unsatu- matter, nitrogen, potassium content and so on, for another, ration observed in TTG is mainly due to the significantly the characteristics of plateau temperate monsoon arid

high proportion（49.6％）of palmitic（C16:0）and oleic acid climate, precipitation deficit also beneficial to maintain the －1 （C18:1）. The peak detected in the range of 3006–3025 cm minerals. While the minerals of Ca, Fe, Mg, Zn detected the is mainly due to high concentration of docosahexaenoic lowest concentration in Lhasa, it was attribute to the local acid（DHA）and eicosapentaenoic acid（EPA）, which were feature. One side, Lhasa located in the southeast of Tibetan found in most fish and marine oils18）. region, which the altitude is over 3650 meters, the plateau In the low-wavenumber region, the prominent peak at temperate semi-arid monsoon climate, sandy loam, the low 1711 cm－1 was detected, which is mainly attributed to the concentration of organic matter lead to the content of min- C＝O carbonyl stretching of the ester functional groups of erals is unstable22）. lipids and fatty acids19）, and the intensity of this peak can The TTG samples subjected to multivariate analysis of measure FFAs in the oils and be used to determine the variance（MANOVA）shows that the minerals in different al- degree of oxidation. Other strong peaks present at 1459 titude level showed the significant variation（p＜0.05）. The and 1120 cm－1 which mainly caused by diacylglycerol results of the minerals based on the altitude level are visu- esters in TTG. Specifically, the peak at 1160 cm－1 repre- alized in the linear discriminant analysis（LDA）plot（Fig.

sents combined asymmetric stretching of（CH3-）and（C-O-） 3a）, the first discriminant function accounted for 72.4％ of bonds. Besides, the distinct peak at 968 cm－1 directly show total variance and the second accounted for 27.6％. All of trans-fats（TFAs）were found in TTG. As a result, FTIR them account for 100％ of total variance, which is very sat- technique is an effective method to evaluate the quality of isfactory. It is shown that there is an adequate separation lipids or oils. in the three altitude levels（A1 to A3）. More specifically, the first discriminant function differentiates the A1（＜3000 3.4 Minerals analysis m）, from A2（4010-3200 m）, the second discriminant func- As shown in Table A.1, the mineral elements in TTG are tion differentiates A2（4010-3200 m）from A3（5165-4010 present in the following order: Ca＞Na＞Mg＞Fe＞Al, m）. As shown in Fig. 4a, calcium is the dominant minerals similar results were reported by Cao et al.20）, moreover, the in TTG, in the altitude range of 3000 m to 5000 m, the content of minerals also showed the significant regional content of Ca increased with altitude level higher signifi- variations（p＜0.05）. For the macroelements of Ca, Mg, Na. cantly, the high concentration of organic matter, magne- Calcium, as the largest content of macroelement, detected sium and phosphorus in soil may explain the high level of in TTG shows the significant regional variations（p＜0.05）, calcium, related to a number of factors such as weather with averages of 497.79 mg /L in Ngari, 265.45 mg /L in conditions, botanical composition of pasture and moisture Shigatse, 298.56 mg /L in Sannan, 386.34 mg /L in Qamdo, content23）. and the level of Ca decreased as the geographical location The TTG samples subjected to MANOVA shows that the from west to east, other minerals such as Al, Mg, Fe, Zn minerals in different regions also showed the significant present the similar tendency. Magnesium which belongs to variation（p＜0.05）. The results of the minerals based on

830 J. Oleo Sci. 68, (9) 827-835 (2019) Characteristics of Traditional Tibetan Ghee

Fig. 3 Linear discriminant analysis based on the minerals of TTG.（ a）Classification of minerals based on the altitude level. （b）Classification of minerals based on the region.（ c）Principal component analysis loading plot based on the minerals in TTG. the regions are visualised in the LDA plot（Fig. 3b）, the （7.54）is lower than western Tibet（8.14）. While in western first discriminant function accounted for 80.5％ of total Tibet, the local soil is over 60.11％, with the high degree of variance and the second accounted for 19.5％, which is sand, moreover, the coverage of vegetation is deficient（＜ very satisfactory. It is shown that there is an adequate sep- 28.90％）, with a small number of organic matter（＜2％） aration of the three groups（B1 to B3）. More specifically, and nitrogen（＜0.21％）, so the gap of minerals in different the first discriminant function differentiates B1（Ngari, regions is mainly attribute to the parameters such as Lhasa）from B3（Qamdo, Nagqu）, the second discriminant organic matter, pH, agrotype and botanical composition of function differentiates B1（Ngari, Lhasa）from B2（Ngari, pasture25）. Shigatse）, the minerals classification based on the regions is mainly attributed to the significant variations in the 3.5 Fatty acid composition analysis physicochemical properties of soil in Tibet24）. In eastern Fatty acid composition for Tibetan ghee of the different Tibet, the the average vegetation coverage is over 68.31％, seven pasture sites is shown in Table A. 2, the total SFAs and the local soil which composed of light and middle loam was 40.99 g/100 g, which is substantially lower than that of riched in organic matter（＞9.16％）, moreover, the high African ghee（46.01％）. Caprylic acid（C8:0）showed the sig- content of moisture（9.16％）and nitrogen（0.42％）in local nificant regional variations（p＜0.05）, with averages of 0.64 soil is beneficial to maintain the minerals, the pH value g/100 g caprylic acid in Nagqu, 0.54 g/100 g in Ngari, 0.51

831 J. Oleo Sci. 68, (9) 827-835 (2019) B. Jing, W. Chen, M. Wang et al.

Fig. 4 Linear discriminant analysis based on the fatty acid composition of TTG.（ a）Classification of fatty acid composition based on the altitude level.（ b）Classification of fatty acid composition based on the region.（ c）Principal component analysis loading plot based on the fatty acids in TTG.

g/100 g in Shigatse, 0.49 g/100 g in Sannan, 0.48 g/100 g in g/100 g）are the dominant fatty acids in TTG. As shown in

Qamdo, and the caprylic acid also varied significantly with Table A. 2, palmitic acid（C16:0）showed the significant re-

the altitude（p＜0.05）. However, the butyric acid（C4:0）and gional variations（p＜0.05）, the level of C16:0 increased as

capric acid（C10:0）did not show significant difference（p＞ the altitude increased, and the highest concentration（27.67

0.05）, with contents lower than that of Nepalese ghee（C4:0 g/100 g）of that was found in Nagqu. Sawaya et al. also re- 3） 1.67 g/100 g, C8:0 0.69 g/100 g, C10:0 2.41 g/100 g）. ported C16:0（27.61–30.5％）and C18:1（19.60–30.11％）are the

Myristic acid（C14:0）（ 3.92-9.85 g/100 g）was lower than predominant fatty acids in ghee and butter from goat’s and 26） that of Indian ghee（11.60 g/100 g）, Nepalese cheese sheep’s milk . The content of palmitic acid（C16:0）is in the （6.71％）, cow milk（10.01 g/100 g）, and yak milk（8.1 g/100 same range of those reported for Nepalese ghee（23.3 g/100 21） g） . Furthermore, the myristic acid（C14:0）showed the sig- g）, French butter（22.27-26.24 g/100 g）, and Holstein ghee nificant variation with the altitude（p＜0.05）, calculating （25.25-25.43 g/100 g）27）; however, it is lower than that in

the averages of 3.94, 4.86, 4.15 g/100 g myristic acid（C14:0） traditional Tunisian butter（32.04％）, Indian ghee（34.09

（respectivly in the altitude of 4900 m, 4500 m and 3400 m） g/100 g）and Aba ghee（28.10％）. Oleic acid（C18:1）was de- in TTG. tected in the range of 22.04-27.85 g/100 g, which is slightly Palmitic（22.19-28.04 g/100 g）and oleic acid（22.04-27.64 higher than that of Holstein ghee（16.98 g/100 g）, goat milk

832 J. Oleo Sci. 68, (9) 827-835 (2019) Characteristics of Traditional Tibetan Ghee

（19.9 g/100 g）, yak milk（20.4 g/100 g）, and Eastern Anato- pasture in different altitudes level show the remarkable lian ghee（15.22-24.02％）. In Switzerland, Collomb et al. effect on milk fatty acid composition34）. reported an average value of 0.7 g CLA/100 g milk fat28）. The TTG samples subjected to MANOVA analysis shows

Linoleic acid（C18:2）n-6 amounts slightly varied depending that the fatty acids in different regions showed the signifi- on region from 0.29 to 1.41 g/100 g. These values are in the cant variation（p＜0.05）, the results of the fatty acids con- same range of the calculated content29）, with the tiny tents based on the region are visualised in the LDA plot

content of linoleic acid（C18:2）, it shows the insignificant （Fig. 4b）. the first discriminant function accounted for variations in altitude and region（p＞0.05）. 78.1％ of total variance and the second accounted for Monosaturated fatty acids（MUFAs）（ oleic and erucylace- 21.9％, both of them account for 100％ of total variance tic acid）account for 34.21g/100 g. These values are higher and separated into three distinct groups（D1 to D3, Fig. than that of Indian ghee（27.32 g/100 g）, Nepalese cheese 4b）, more specifically, the first discriminant function dif- （32.40 g/100 g）, and traditional Tunisian butter（24.68％）30）. ferentiates D2（Lhasa, Sannan）from D3（Ngari, Shigatse）, Moreover, the health-affecting polyunsaturated fatty acids the second discriminant function differentiates D1（Qamdo, （PUFAs）such as linoleic acid and eicosapentaenoic acid Nagqu）from D2（Lhasa, Sannan）. And the classification （EPA）is abundant, with 2.46 g/100 g31）, which is slightly based on the regions in fatty acids mainly links to the local higher than that of Tunisian butter（2.41 g/100 g）and grass and fodder qualities, and all the factors such as re- Baffalo milk（2.13 g/100 g）; however, it is lower than that of gional weather, cow diet, breeding modes and bovine the Holstein ghee（3.80 g/100 g）, yak milk（4.91 g/100 g）, breed, may modulate the fatty acids contents of TTG.

and Indian ghee（2.36 g/100 g）. The ratio of unsaturated to Figure 4c shows that C18:1 and C16:0 are the dominant fatty saturated（UFA /SFA）fatty acid is also an important criteri- acids in TTG, and the dominant fatty acids are confirmed on to evaluate the quality of TTG, which was approximate- the regional variations, with averages of 48.41 g/100 g ghee ly 0.83. The value of UFA /SFA is higher than that of Indian in D3（Ngari, Shigatse）, 50.21 g/100 g in D2（Lhasa, ghee（0.41）, yak milk（0.78）, cow milk（0.57）, and Tunisian Sannan）, and 50.65 g/100 g in D1（Qamdo, Nagqu）, these butter（0.38）32）. The high level of UFA/SFA indicated the results show that the fatty acids level are higher in eastern fatty acid composition in TTG is abundant and healthy, and regions than in western area in Tibet. As reported, the geo- the health-affecting fatty acids in TTG such as oleic acid graphical characteristics of the regions in Tibet（D1, D2,

（C18:1）, linoleic acid（C18:2）, CLAs（C18:2 cis 9, trans 11）, ara- D3）is significantly different, the local characteristics of D3

chidonic acid（C20:4）and EPA are usually detected in ocean （Ngari, Shigatse）are arid climate, high altitude level, pre- 33） fish oil . Moreover, the PUFAs in TTG（C18:2, CLAs, C20:5） cipitation deficit, and the fodder of yak is mainly composed are in the range of 1.65-2.35 g/100 g. The effect of regional of alpine meadow, upland meadow and brushwood. In D2

variations on the EPA（C20:5）level showed significantly（p＜（Lhasa, Sannan）region, the regional weather belongs to 0.05）, the averages of 0.47 g/100 g in Sannan, 0.39 g/100 g the plateau temperate semi-arid monsoon climate, with in Shigatse, 3.84 g/100 g in Nagqu7）. strong solar radiation, long sunshine time, and there are The TTG samples subjected to MANOVA analysis shows abundant fodders such as shrubs and herbs, many of which that the fatty acids in different altitude level showed the are known for medicinal, aromatic, and nutritive proper- significant variation（p＜0.05）. Moreover, the results of the ties. The regional climate features in D1（Qamdo, Nagqu） fatty acids contents based on the altitude level are visual- region mainly shows the abundant sunshine, low humidity, ised in the LDA plot（Fig. 4a）, the first discriminant func- and rainfall concentration, moreover, the area is covered tion accounted for 73.2％ of total variance and the second with forest and grassland, many of which are rich in abun- accounted for 26.8％, both of them account for 100％ of dant PUFAs35）. The regional effect on the grass pasture total variance, which is very satisfactory. It is shown in Fig. qualities, related to cow diet, ultimately influence the total 4a that there is an adequate separation of the three alti- fatty acid composition of TTG, moreover, the factors such tude levels. More specifically, the first discriminant func- as breeding methods, fodder quality and technology, cow tion differentiates C1（＜3000 m）from C2（4021-3200 m）, breeds may affect fatty acids content of TTG. the second discriminant function differentiates C1（＜3000

m）from C3（5165-4100 m）. Figure 4c shows that the C18:1

and C16:0 belong to the predominant fatty acids, with averages of 50.73 g/100 g ghee in C3（5165-4100 m）, 49.36 4 Conclusion g/100 g in C2（4021-3200 m）, and 50.11 g/100 g in C1（＜ TTG, as the dairy products is nourishing and attractive 3000 m）, it shows that the level of fatty acids is higher in food for consumers, which actually contains some health- highlands than in lowlands, the effect of altitude level on affecting fatty acids. It is important to update food data for botanical composition of pasture, related to yak diets, ulti- physicochemical characteristics and a number of particular mately influence the total fatty acid composition. Collomb fatty acids such as CLAs and EPA. et al. previously also reported the botanical composition of TTG contains high proportion of dry matter, such as fat

833 J. Oleo Sci. 68, (9) 827-835 (2019) B. Jing, W. Chen, M. Wang et al.

（71.68％–93.39％）, protein（0.51％–1.8％）, and minerals Golecky, J.; Martin, B.; Ferlay, A.; Constant, I.; Dela- （97.31–887.87 mg/L）. Quality parameters such as AV and vaud, C.; Hurtand, C. Mineral, vitamin a and fat com- PV are within 0.02–1.30 mg/g and 0.07–5.93 meq/kg, re- position of bulk milk related to european production spectively, the low PV indicates that the low quantities of conditions throughout the year. Dairy Sci. Technol. FFAs are in TTG. As detected by the FTIR spectra, high 96, 715-733（2016）. proportion（49.6％）of UFAs easily lead to generating rancid 5） Silva, C.C.; Silva, S.P.; Prates, J.A.; Bessa, R.J.; Rosa, flavor. The minerals increased with altitude level higher H.J.; Rego, O.A. Physicochemical traits and sensory significantly, and it also showed significant regional varia- quality of commercial butter produced in the Azores. tions. Moreover, TTG is an excellent source of UFAs, espe- Int. Dairy J. 88, 10-17（2019）. cially medium and long chain fatty acids, and the content 6） Samet-Bali, O.; Ayadi, M.A.; Attia, H. Traditional Tuni- of fatty acids varied with altitude level and regions. The sian butter: physicochemical and microbial character- level of fatty acids in highlands is higher than in lowlands, istics and storage stability of the oil fraction. LWT- however, the content of fatty acids in eastern regions is Food Sci. Technol. 42, 899-905（2009）. higher than western area in Tibet. These variations is 7） Kirazci, A.; Javidipour, I. Some chemical and microbio- mainly due to the differences in grass pasture qualities, re- logical properties of ghee produced in Eastern Anato- gional weather, related to the breeding methods, grass and lia. Int. J. Dairy Technol. 61, 300-306（2008）. fodder qualities, yak diets, and so on. 8） Idoui, T.; Benhamada, N.; Leghouchi, E. Microbial quality, physicochemical characteristicsand fatty acid composition of a traditional butter produced from cows’ milk in East Algeria. Grasas Aceites 61, 232-236 Acknowledgement （2010）. The authors would like to thank the National Natural 9） Fındık, O.; Andiç, S. Some chemical and microbiologi- Science Foundation of China（NO.: 31671819）for the finan- cal properties of the butter and the butter oil pro- cial support. duced from the same raw material. LWT-Food Sci. Technol. 86, 233-239（2017）. 10） Sserunjogi, M.L.; Abrahamsen, R.K.; Narvhus, J. A review paper: current knowledge of ghee and related Conflict of Interest products. Int. Dairy J. 8, 677-688（1998）. The authors have declared no conflicts of interest. 11） Maryam, J.; Mahdi, K.; Javad, K. Detection of adultera- tion in Iranian olive oils using instrumental（GC, NMR, DSC）methods. J. Am. Oil Chem. Soc. 86, 103-110 （2009）. Supporting Information 12） Hjorth, J.L.; Miller, R.L.; Woodley, J.M.; Kiil, S. Ther- This material is available free of charge via the Internet modynamic modeling of multi-phase solid–liquid equi- at http://dx.doi.org/jos.68.10.5650/jos.ess19031 libria in industrial-grade oils and fats. J. Am. Oil Chem. Soc. 92, 17-28（2015）. 13） Ferrari, C.; Angiuli, M.; Tombari, E.; Righetti, M.C.; Matteoli, E.; Salvetti, G.; Promoting calorimetry for ol- References ive oil authentication. Thermochimica Acta 459, 1） El-Rahman, A.A.; Shalabi, S.I.; Kilara, A. Chemical and 58-63（2007）. physical properties of anhydrous milk fat, low and very 14） Harrison, P.D.; Smith, K.W.; Bhaggan, K.; Stapley, high melting milk fat fractions obtained by a commer- A.G.F. Image analysis of palm oil crystallisation as ob- cial fractionation process. Milchwissenschaft 53, served by hot stage microscopy. J. Cryst. Growth 444, 77-81（1998）. 28-38（2016）. 2） Neupaney, D.; Kim, J.; Ishioroshi, M.; Samejima, K. 15） dos Santos, M.T.; Viana, I.S.; Ract, J.N.R.; Le Roux, Study on some functional and compositional proper- G.A.C. Thermal properties of palm stearin, canola oil ties of yak butter lipid. J. Anim. Sci. 74, 391-397 and fully hydrogenated soybean oil blends: coupling （2015）. experiments and modeling. J. Food Eng. 185, 17-25 3） Orrashid, M.M.; Odongo, N.E.; Subedi, B.; Karki, P.; （2016）. Mcbride, B.W. Fatty acid composition of yak（Bos 16） Guillén, M.D.; Cabo, N. Infrared spectroscopy in the grunniens）cheese including conjugated linoleic acid study of edible oils and fats. J. Sci. Food Agric. 75, and trans-18:1 fatty acids. J. Agric. Food Chem. 56, 1-11（2016）. 1654-1660（2008）. 17） Vongsvivut, J.; Miller, M.R.; Mcnaughton, D.; Heraud, P.; 4） Chassaing, C.; Sibra, C.; Verbic, J.; Harstad, O.M.; Barrow, C.J. Rapid discrimination and determination

834 J. Oleo Sci. 68, (9) 827-835 (2019) Characteristics of Traditional Tibetan Ghee

of polyunsaturated fatty acid composition in marine Escher, F.; Schlichtherlecerny, H. Determination of oils by FTIR spectroscopy and multivariate data analy- storage stability of butter enriched with unsaturated sis. Food Bioprocess Tech. 7, 2410-2422（2014）. fatty acids /conjugated linoleic acids（UFA/CLA）using 18） Rohman, A.; Man, Y.B.C. Application of FTIR spectros- instrumental and sensory methods. Int. Dairy J. 18, copy for monitoring the stabilities of selected vegeta- 983-993（2008）. ble oils during thermal oxidation. Int. J. Food Prop. 28） Collomb, M.; Bütikofer, U.; Sieber, R.; Bosset, O.; Jean- 16, 1594-1603（2013）. gros, B. Conjugated linoleic acid and trans fatty acid 19） Zhang, Q.; Liu, C.; Sun, Z.; Hu, X.; Shen, Q.; Wu, J. Au- composition of cows’ milk fat produced in lowlands thentication of edible vegetable oils adulterated with and highlands. J. Dairy Res. 68, 519-523（2001）. used frying oil by Fourier transform infrared spectros- 29） Fei, T.; Peng, W.; Lin, Y.; Ying, M.; Li, D. Quantification copy. Food Chem. 132, 1607-1613（2012）. of fatty acids in human, cow, buffalo, goat, yak, and 20） Cao, Y.; Yao, X.; Ren, H.; Wang, K. Determination of camel milk using an improved one-step GC-FID meth- fatty acid composition and metallic element content of od. Food Anal. Method. 10, 2881-2891（2017）.

four camellia, species used for edible oil extraction in 30） Precht, D.; Molkentin, J. C18:1, c18:2 and c18:3 trans and China. J. Verbrauch. Lebensm. 12, 1-5（2017）. cis fatty acid isomers including conjugated cis delta 9, 21） Niu, Y.J.; Yang, S.W.; Wang, G.Z.; Liu, L.; Du, G.Z.; Hua, trans delta 11 linoleic acid（cla）as well as total fat L.M. Relation between species distribution of plant composition of german human milk lipids. Nahrung community and soil factors under grazing in alpine 43, 233-244（1999）. meadow. J. Appl. Ecol. 28, 3891-3898（2017）. 31） Ixtaina, V.Y.; Martínez, M.; Spotorno, V.; Mateo, C.M.; 22） Wang, Z.; Zhang, Y.; Zhang, H.; Ding, M.; Lin, X. Major Maestri, D.M.; Diehl, B.W.K.; Tomás, M.C. Character- soil element（Ca, Mg, K, Na, Al, Fe）distribution along ization of chia seed oils obtained by pressing and sol- the Qinghai-Tibet Railway. in AGU Fall Meeting, Ab- vent extraction. J. Food Compos. Anal. 24, 166-174 stracts（2011）. （2011）. 23） Tipper, E.T.; Galy, A.; Bickle, M.J. Calcium and magne- 32） Ledoux, M.; Chardigny, J.M.; Darbois, M.; Soustre, Y.; sium isotope systematics in rivers draining the himala- Sebedio, J.L.; Laloux, L. Fatty acid composition of ya-tibetan-plateau region: lithological or fractionation french butters, with special emphasis on conjugated control? Geochim. Cosmochim. Acta 72, 1057-1075 linoleic acid（CLA）isomers. J. Food Compos. Anal. （2008）. 18, 409-425（2005）. 24） Wang, C-T.; Long, R-J.; Wang, G-X.; Liu, W.; Wang, 33） Addis, M.; Cabiddu, A.; Pinna, G.; Decandia, M.; Pired- Q-L.; Zhang, L.; Wu, P-F. Relationship between plant da, G.; Pirisi, A.; Molle, G. Milk and cheese fatty acid communities, characters, soil physical and chemical composition in sheep fed mediterranean forages with properties, and soil microbiology in alpine meadows. reference to conjugated linoleic acid cis-9, trans-11. J. Acta Prataculturae Sinica 19（6）, 25-34（2010）. Dairy Sci. 88, 3443-3454（2005）. 25） Sawaya, W.N.; Khan, P.; Al-Shalhat, A.F. Physical and 34） Collomb, M. Composition of fatty acids in cow’s milk chemical characteristics of ghee and butter from goat’s fat produced in the lowlands, mountains and highlands and sheep’s milk. Food Chem. 14, 227-232（1984）. of Switzerland using high-resolution gas chromatogra- 26） Zhang, X.; Yu, Y.; Jiang, Z. The interaction between phy. Int. Dairy J. 12, 649-659（2002）. poisonous plants and soil quality in response to grass- 35） Dorea, J.R.R.; French, E.A.; Armentano, L.E. Use of land degradation in the alpine region of the Qinghai-ti- milk fatty acids to estimate plasma nonesterified fatty betan plateau. Plant Ecol. 215, 809-819（2014）. acid concentrations as an indicator of animal energy 27） Mallia, S.; Piccinali, P.; Rehberger, B.; Badertscher, R.; balance. J. Dairy Sci. 100, 6164-6176（2017）.

835 J. Oleo Sci. 68, (9) 827-835 (2019)