Samples of animal fats namely lard, beef fat, mutton fat, and chicken fat were extracted by rendering three samples of adipose tissues of animals collected from local slaughter houses located in the Bangi, Serdang and Kajang areas of Malaysia according to the method reported previously by Marikkar et al. (2001). Analytical grade chemicals of glycerol, hexane, diethyl ether, chloroform and sodium hydroxide pellets were obtained from Merck Chemicals, Germany. A set of FAME standard comprising 37 fatty acids (C4 to C24) and a monoacylglycerol stock solution were purchased from Sigma–Aldrich Chemicals (Deisenhofen, Germany).

Chemical glycerolysis of animal fats was performed in triplicate according to the procedure described by Indrasti et al. (2010). For the glycerolysis reaction, a 35-g fat sample was mixed with 15 g of a glycerine solution and a 0.2 g portion of sodium hydroxide. The mixture was then heated at 250 °C with vigorous mixing for 60 min. The separation of glycerolysis products into MAG and DAG was carried out according to the AOCS method Cd 11c- 93 (AOCS, 2007) using a glass column filled with Davison 923 type silica gel (Sigma Aldrich). Thin layer chromatography was run subsequently to verify the purity of the fractions collected from column chromatography.

About 0.2 µg of partial acylglycerol of individual animal fat was weighed and loaded into a clean tin capsule to determine their δ¹³C values. The capsules containing samples were placed onto the auto-sampler system of an elemental analyzer (Europa Scientific, UK) for burning in an O₂ atmosphere of the combustion CuO tube with its temperature set at 960 °C. Combustion gases were eluted through a reduction column by a stream of He gas and passed into the gas chromatograph where CO₂, still in the He stream, was separated from the other gases. The gas stream was then entered into the IRMS system (Sercon Ltd., Crewe, U.K.) where the CO₂ gas was analyzed by comparison with NBS-22 reference material (with a δ¹³C value of −30.03‰). During every batch of analyses, an empty tin capsule was analyzed as the blank to check the background (Liu et al., 2007). The results were compared to Vienna Pee Dee Belemnite (V-PDB). The isotopic values were calculated against the international isotope reference standards: NBS-22 (International Atomic Energy Agency, Vienna, Austria) for 13C/12C measurements (Osorio et al., 2011).

A 50-mg portion of partial acylglycerol was weighed into a 20-mL test tube (with screw cap). After adding a 2 ml portion of 2 N sodium hydroxide in methanol, the sample tube was closed and heated at 80 °C for 1 hour. After allowing the tube and its contents to set for a few minutes, a 2 mL portion of 25% borontrifluoride solution in methanol was added. The tube was closed and heated again for 1 hour at 80 °C. Subsequently, 5 mL portions of water and hexane were added into this. The contents of the tube were shaken well and allowed to undergo phase separation. The clear supernatant of the solution was transferred to a 2-mL auto-sampler vial (AOAC, 2007).

The top hexane layer of the FAME solution was injected onto an Agilent 6890N gas chromatograph (Agillent Technologies, Singapore) equipped with a polar capillary column RTX-5 (0.32 mm internal diameter, 30 m length and 0.25 µm film thickness; Restex Corp., Bellefonte, PA) and a Flame Ionization Detector (FID). Split injection was conducted with a split ratio of 58:1 using nitrogen as carrier gas at a flow-rate of 1.00 mL·min⁻¹. The temperature of the column was 50 °C (for 1 min), and programmed to increase to 200 °C at 8 °C·min⁻¹. The temperatures of the injector and detector were maintained at 200 °C (Yanty et al., 2011).

Data were statistically analyzed by one-way analysis of variance (ANOVA) using MINITAB (version 14) statistical package at the 0.05 probability level. For the grouping and classification of models, PCA was carried out using Unscrambler 9.7 (Camo, USA) software.

The δ¹³C values of MAG and DAG of lard, chicken fat, beef fat and mutton fat are compared as shown in Table 1. The δ¹³C values of MAG and DAG of lard, beef fat, chicken fat and mutton fat have not been investigated previously; therefore there are hardly any reports to compare the δ¹³C value. The highest δ¹³C value of MAG was found in chicken fat (−20.3‰) while the lowest value was found for mutton fat (−31.9‰). The δ¹³C values of MAG of lard (−22.3‰) and beef fat (−25.0‰) were within the range of these two δ¹³C values. In the case of DAG, the highest δ¹³C value of DAG was found for chicken fat (−19.2‰) while the lowest value was found for mutton fat (−32.2‰). The δ¹³C values of the DAGs of lard (−22.2‰) and beef fat (−24.3‰) were within the range of these two extremes of the mean δ¹³C values. The statistical analysis of the data from the present study suggested that the determination of δ¹³C value for MAG and DAG can be a useful tool since the δ¹³C value of lard (−22.9‰) was significantly (p < 0.05) different from those of beef fat (−25.0‰), chicken fat (−20.3‰), and mutton fat (−31.9‰) (Table 1). The observed variation in the δ¹³C values of animal fat could be attributed to their species difference (Osorio et al., 2011), genetic factors (Wood et al., 2008) as well as the diet fed to the animals (Bojlul et al., 2007). According to previous investigators, the variation in the δ¹³C values of oils and fats originating from different plant sources are due to isotopic fractionation during physical, chemical and biological processes in plants (Kelly and Rhodes, 2002).

The fatty acid distributional pattern of MAG derived from lard, chicken fat, beef fat and mutton fat are compared as shown in Table 2. The most dominant fatty acid of MAG derived from lard was oleic (38.75%), followed by palmitic (28.16%) and linoleic (21.55%). According to previous reports, this was the distributional order of the major fatty acids of lard (Marikkar & Yanty, 2014; Yanty et al., 2011). The MAG of chicken fat was also found to have oleic acid (45.94%) as the most dominant fatty acid, followed by palmitic (30.99%) and linoleic (8.97%) acids. The MAG of chicken fat differed from that of lard by having a small amount of palmitoleic acid (6.01%) in addition to differences in the proportions of other fatty acids. The MAGs derived from beef and mutton fats differed from those of lard and chicken fat by having stearic (31.31–35.34%) as the major fatty acid (Table 2). For this reason, the MAG of beef fat and mutton fat displayed lower contents of unsaturated fatty acids (30.61 to 39.72%) than saturated fatty acids (60.27 to 69.39%). In fact, there were differences between the MAG of beef fat and mutton fats with regard to the occurrence of myristic and linoleic acids. While some amounts of myristic and linoleic acids were present in the MAG of beef fat, they were totally absent in the MAG of mutton fat. These differences were mainly due to the diverse pattern of distribution of individual fatty acids among animal fats as reported previously (Naquiya et al., 2013). When these data were subjected to an appropriate multivariate data analysis technique, a clear differentiation can be reached. PCA, for instance, is a way of identifying patterns in data, and expressing the data in such a way as to emphasize their similarities and differences (Shin et al., 2010).

As fatty acids, namely myristic (C14:0), palmitic (C16:0), palmitoleic (C16:1), stearic (C18:0), oleic (C18:1) and linoleic (C18:2) acids were found to occur in varying amounts in the MAG of all four animal fats, they were used as independent variables in the PCA procedure. According to Kamal-Eldin and Anderson (1997), visualization of groups by various plotting systems becomes possible if data of at least three measured variables are available. The score plot of the fatty acids of the MAG derived from the MAGs of four animal fats as shown in Figure 1 represents the projection of samples defined by principle component 1 (PC1) and principle component 2 (PC2). PC1 is the linear combination of variables that explains the highest variation among the samples, while PC2 is orthogonal to PC1 and exhibited the second largest variation (Cordella et al., 2003). The score plot projected on PC1 described 89% of the variation while PC2 accounted for 8% of the variation, making up 97% of variance for PC1 and PC2. According to the group separation illustrated in Figure 1, the MAG of mutton fat and beef fat were located in the upper right quadrant and lower right quadrant, respectively. Their positions were well differentiated from the MAGs of lard and chicken fat, which were located in the lower left quadrant, and upper left quadrant, respectively. However, the separation between the MAGs of lard and chicken fat was not remarkable due to their close proximity to each other. Fatty acid variables having high influence on the group separation of the samples in the score plot could be traced from the analysis of the loading plot. As explained by Cordella et al. (2003), a variable which is farther from the origin of the axis contributes most to the variation in the statistical model generated by PCA. According to the loading plot shown in Figure 2, out of the six fatty acid variables, stearic, oleic and linoleic acids were the most discriminating variables that influenced the group separation into four different clusters.

The data presented in Table 2 compares the fatty acid distributional pattern of DAG derived from lard, chicken fat, beef fat and mutton fat. As noted before with MAG, the most dominant fatty acid of DAG derived from lard was oleic acid (40.29%), followed by palmitic (31.35%) and stearic (14.92%). Similarly, the DAG of chicken fat was also found to have oleic acid (47.24%) as the most dominant fatty acid, followed by palmitic (31.36%) and stearic (7.85%) acids. The DAG of chicken fat differed from that of lard by having differences in the proportions of several fatty acids. In addition, the occurrence of a small amount of arachidic acid in the DAG of chicken fat was noteworthy. DAG derived from beef and mutton fats differed considerably from those of lard and chicken fat by having stearic (36.49–38.57%) as the major fatty acid (Table 2). As noted before in the case of MAG fatty acids, the DAG of beef fat and mutton fat also displayed lower contents of unsaturated fatty acids (30.99 to 36.13%) than saturated fatty acids (63.87 to 69.02%). There were differences between the MAG of beef and mutton fats with regard to the occurrence of myristic, palmitoleic, and linoleic acids as these three fatty acids were totally absent in the MAG of mutton fat. Despite the observed differences, a clear differentiation among the DAG of different animal fats can be reached when these data are subjected to a multivariate data analysis technique such as PCA.

As listed in Table 2, fatty acids, namely myristic (14:0), palmitic (C16:0), palmitoleic (C16:1), stearic (C18:0), oleic (C18:1) and linoleic (C18:2) acids were found to occur in variable amounts in all the DAG derived from animal fats. When PCA was performed using these fatty acids as variables, a classification of DAG originating from different animal fats was established. The score plot of the fatty acids of DAG derived from animal fats shown in Figure 3 represent the projection of samples defined by principle component 1 (PC1) describing 87% of the variation while PC2 accounted for 9% of the variation, making up 96% of variance explained for the PC1 and PC2. According to the group separation illustrated in Figure 3, the DAG of lard and chicken fat were located in the lower left quadrant, and upper left quadrant, respectively. Their positions were well differentiated from the DAG of mutton fat and beef fat, which were located in the upper right quadrant and lower right quadrant, respectively. Fatty acid variables having strong influence on the group separation of the samples in the score plot could be traced from the analysis of the loading plot as presented in Figure 4. As explained by Cordella et al. (2003), a variable which is farther from the origin of axis contributes most to the variation in the statistical model generated by PCA. According to the loading plot in Figure 4, out of the six fatty acids, stearic, oleic and linoleic acids were the most discriminating variables that influence the group separation into four different clusters.