Preparation of Human Milk Fat Substitutes from Lard by Lipase‐Catalyzed Interesterification Based on Triacylglycerol profiles

Human milk fat substitutes (HMFSs) with triacylglycerol profiles highly similar to those of human milk fat (HMF) were prepared from lard by physical blending followed by enzymatic interesterification. Based on the fatty acid profiles of HMF, different vegetable and single‐cell oils were selected and added to the lard. Blend ratios were calculated based on established physical blending models. The blended oils were then enzymatically interesterified using a 1,3‐regiospecific lipase, Lipozyme RM IM (RML from Rhizomucor miehei immobilized on Duolite ES562; Novozymes A/S, Bagsværd, Denmark), to approximate HMF triacylglycerol (TAG) profiles, particularly with respect to the distribution of palmitic acid in the sn?2 position. The optimized blending ratios were determined to be: lard:sunflower oil:canola oil:palm kernel oil:palm oil:algal oil:microbial oil = 1.00:0.10:0.50:0.13:0.12:0.02:0.02. The optimized reaction conditions were determined to be: enzyme load of 11 wt%, temperature of 60 °C, water content of 3.5 wt%, and reaction time of 3 hours. The resulting product was evaluated for total and sn?2 fatty acids, polyunsaturated fatty acids, and TAG composition. A high degree of similarity was obtained, indicating the great potential of the product as a fat alternative for use in infant formulas.     

Lipase-Catalyzed Synthesis of Human Milk Fat Substitutes from Palm Stearin in a Continuous Packed Bed Reactor

Structured lipids (SL) with similar fatty acid (FA) composition and distribution to human milk fat (HMF) were synthesized by lipase-catalyzed acidolysis of chemically interesterified palm stearin (IV = 35.6) with mixed FA of stearic acid and myristic acid and FA from rapeseed oil, sunflower oil, and palm kernel oil in a continuous packed bed reactor. Response surface methodology (RSM) was used to optimize the reaction system with three selected parameters, namely residence time, temperature, and substrate molar ratio. The best-fitting quadratic models were obtained for the contents of palmitic acid (PA) and PA at the sn-2 position (sn-2 PA) by multiple regressions and the determination coefficient (R ²) values for the models were 0.9886 and 0.9799, respectively. The optimal conditions generated from the models were as follows: residence time, 2.7 h; temperature, 58 °C; substrate molar ratio, 9.5 mol/mol. Under these conditions, the contents of PA and sn-2 PA were 28.8 and 53.2%, respectively, and other FA observed in the experiments were all within the range of corresponding FA of HMF. The similarity of the product obtained to HMF was evaluated by the cited model. The scores for total and sn-2 FA of the product were 45.2 and 38.4, respectively, and the total score for the product was 83.6, which indicated a high degree of similarity of the product to HMF.     

The influence of chemical interesterification on physicochemical properties of complex fat systems 1. Melting and crystallization

The effects of blending palm oil (PO) with soybean oil (SBO) and lard with canola oil, and subsequent chemical interesterification (CIE), on their melting and crystallization behavior were investigated. Lard underwent larger CIE-induced changes in triacylglycerol (TAG) composition than palm oil. Within 30 min to 1 h of CIE, changes in TAG profile appeared complete for both lard and PO. PO had a solid fat content (SFC) of ∼68% at 0°C, which diminished by ∼30% between 10 and 20°C. Dilution with SBO gradually lowered the initial SFC. CIE linearized the melting profile of all palm oil-soybean oil (POSBO) blends between 5 and 40°C. Lard SFC followed an entirely different trend. The melting behavior of lard and lard-canola oil (LCO) blends in the 0–40°C range was linear. CIE led to more abrupt melting for all LCO blends. Both systems displayed monotectic behavior. CIE increased the DP of POSBO blends with ≥80% PO in the blend and lowered that of blends with ≤70% PO. All CIE LCO blends had a slightly lower DP vis-à-vis their noninteresterified counterparts.     

Interesterification of Lard and Soybean Oil Blends Catalyzed by Immobilized Lipase in a Continuous Packed Bed Reactor

Structured lipids (SL), formulated by blends of lard and soybean oil in different ratios, were subjected to continuous enzymatic interesterification catalyzed by an immobilized lipase from Thermomyces lanuginosus (Lipozyme TL IM) in a continuous packed bed reactor. The original and interesterified blends were examined for fatty acid and triacylglycerol composition, regiospecific distribution, and solid fat content. Blends of lard and soybean oil in the proportions 80:20 and 70:30 (w/w), respectively, demonstrated a fatty acid composition, and proportions of polyunsaturated/saturated fatty acids (PUFA/SFA) and monounsaturated/polyunsaturated fatty acids (MUFA/PUFA), that are appropriate for the formulation of pediatric products. These same blends were suited for this purpose after interesterification because their sn-2 positions were occupied by saturated fatty acids (52.5 and 45.4%, respectively), while unsaturated fatty acids predominantly occupied sn-1,3 positions, akin to human milk fat. Interesterification caused rearrangement of triacylglycerol species.     

Biotechnological and Novel Approaches for Designing Structured Lipids Intended for Infant Nutrition

Human milk fat (HMF) is a perfect nutritional source that includes all the required ingredients which are necessary for the growth of infants up to 6 months. Although its composition may differ among mothers or during lactation stage, its unique triacylglycerol (TAG) structure remains constant which is characterized by the presence of palmitic acid (PA) at the sn‐2 position. Previous reports provided convincing information of higher PA and calcium absorption and efficient use of dietary energy when at this specific position in the TAG moiety than when PA is at the sn‐1,3 positions. During the design of structured lipids (SLs) intended for infant nutrition, this unique property is taken into consideration. Human milk fat substitutes (HMFS) enriched with important fatty acids such as omega‐3 and omega‐6 fatty acids are intended to better mimic the functions of HMF as well as provide associated health benefits. The use of microencapsulation technology and novel technologies such as ultrasound technology in conjunction with SL production and enzyme‐catalyzed reactions are evolving and ongoing issues in infant formula production. Therefore, further studies should be directed towards new process improvements in order to increase the functional properties and oxidative stabilities of HMFS. Novel technologies in lipid biotechnology related to HMFS preparation should also be explored.     

Application of palm olein in the production of zero‐trans Iranian vanaspati through enzymatic interesterification

The utilization of palm olein in the production of zero‐trans Iranian vanaspati through enzymatic interesterification was studied. Vanaspati fat was made from ternary blends of palm olein (POL), low‐erucic acid rapeseed oil (RSO) and sunflower oil (SFO) through direct interesterification of the blends or by blending interesterified POL with RSO and SFO. The slip melting point (SMP), the solid fat content (SFC) at 10–40 °C, the carbon number (CN) triacylglycerol (TAG) composition, the induction period (IP) of oxidation at 120 °C (IP₁₂₀) and the IP of crystallization at 20 °C of the final products and non‐interesterified blends were evaluated. Results indicated that all the final products had higher SMP, SFC, IP of crystallization and CN 48 TAG (trisaturated TAG), and lower IP₁₂₀, than their non‐interesterified blends. However, SMP, SFC, IP₁₂₀, IP of crystallization and CN 48 TAG were higher for fats prepared by blending interesterified POL with RSO and SFO. A comparison between the SFC at 20–30 °C of the final products and those of a commercial low‐trans Iranian vanaspati showed that the least saturated fatty acid content necessary to achieve a zero‐trans fat suitable for use as Iranian vanaspati was 37.2% for directly interesterified blends and 28.8% for fats prepared by blending interesterified POL with liquid oils.     

Operational stability of Thermomyces lanuginosa lipase during interesterification of fat in continuous packed‐bed reactors

The operational stability of a commercial immobilized lipase from Thermomyces lanuginosa (“Lipozyme TL IM”) during the interesterification of two fat blends, in solvent‐free media, in a continuous packed‐bed reactor, was investigated. Blend A was a mixture of palm stearin (POS), palm kernel oil (PK) and sunflower oil (55 : 25 : 20, wt‐%) and blend B was formed by POS, PK and a concentrate of triacylglycerols rich in n‐3 polyunsaturated fatty acids (PUFA) (55 : 35 : 10, wt‐%). The bioreactor operated continuously at 70 °C, for 580 h (blend A) and 390 h (blend B), at a residence time of 15 min. Biocatalyst activity was evaluated in terms of the decrease of the solid fat content at 35 °C of the blends, which is a key parameter in margarine manufacture. The inactivation profile of the biocatalyst could be well described by the first‐order deactivation model: Half‐lives of 135 h and 77 h were estimated when fat blends A and B, respectively, were used. Higher levels of PUFA in blend B, which are rather prone to oxidation, may explain the lower lipase stability when this mixture was used. The free fatty acid content of the interesterified blends decreased to about 1% during the first day of operation, remaining constant thereafter.     

Interesterification of sesame oil and a fully hydrogenated fat using an immobilized lipase catalyst in both batch and continuous‐flow processes

Lipase‐mediated interesterification of sesame oil and a fully hydrogenated soybean oil was studied at 70 °C in both a batch reactor (BR) and a continuous‐flow packed‐bed reactor (PBR) using four different initial weight ratios of substrates (90 : 10, 80 : 20, 70 : 30 and 60 : 40) with Lipozyme TL IM (Thermomyces lanuginosa) as the biocatalyst. Reaction rates were determined by following the dependence of the profile of the product triacylglycerols (TAG) on the reaction time (BR) or the space time (PBR) via RP‐HPLC‐ELSD. Product TAG identities were confirmed by HPLC‐APCI‐MS. Primary differences between the performances of the two reactors were the maximum level of net hydrolysis (ca. 3 and 10 wt‐% lower acylglycerols at equilibrium for the PBR and BR, respectively), the time or space time required to approach quasi‐equilibrium conditions, and less migration of acyl groups in the PBR trials. For the BR trials, quasi‐equilibrium conditions were approached in 4–6 h, while for the PBR trials short space times (15 min to 2 h) were sufficient to produce effluent compositions similar to equilibrium BR compositions. The predominant TAG families formed by interesterification were LLS, PSO, PSL, SSL, and SSO (L = linoleic; S = stearic; P = palmitic; O = oleic). Oxidative stabilities, melting profiles and solid fat contents were determined for selected reaction products.     

Process optimization of simultaneous enzymatic production of 1,3-dioleoyl-2-palmitoylglycerol and 1-oleoyl-2-palmitoyl-3-linoleoylglycerol

1,3-dioleoyl-2-palmitoylglycerol (OPO), a structured triacylglycerols (TAG) containing palmitic acid at sn-2 position, is widely used as breast milk fat substitutes in infant formula. Interestingly, the main component of breast milk fat in Western countries is OPO while Chinese breast milk contains both 1-oleoyl-2-palmitoyl-3-linoleoylglycerol (OPL) and OPO. Thus, OPL is additionally essential for Chinese infants. In the present study, we developed a process for simultaneous production of OPL and OPO. According to the fatty acid (FA) composition of Chinese breast milk, soy oil, palm kernel stearin, and palm stearin were selected as the base oils for random interesterification, and the appropriate ratio was determined as 2:3:15. Then interesterification at sn-1,3 position was conducted, four reaction parameters were examined, where 1.6 molar ratio of oleic acid to linoleic acid (OLA/LNA), 10% enzyme quantity, TAG to FA ratio of 1:8 and reaction temperature of 56°C were ultimately established as the optimal conditions to achieve the target of the present study, and the OPO and OPL contents in the final products were 23.1% and 26.2%, respectively. Finally, molecular distillation was applied to purify above achieved breast milk fat substrates. Higher oil dropping speed, oil temperature, and agitation speed were favorable conditions, and the lowest acid value that achieved in the present study was 3.46 mgKOH/g. Overall, we established an efficient preliminary process for the simultaneous production of OPO and OPL, which could provide an important reference for commercial production of breast milk fat substrates for Chinese infants.     

Enzymatic Interesterification of Extra Virgin Olive Oil with a Fully Hydrogenated Fat: Characterization of the Reaction and Its Products

The lipase-catalyzed interesterification of extra virgin olive oil (EVOO) and fully hydrogenated palm oil (FHPO) was studied in a batch reactor operating at 75 °C. The compositions of the semi-solid fat products depend on the reaction conditions and the initial ratio of EVOO to FHPO. The dependence of the quasi-equilibrium product TAG profile on the reaction time was determined for initial weight ratios of EVOO to FHPO from 80:20 to 20:80. Lipozyme TL IM, Lipozyme RM IM and Novozym 435 were employed as biocatalysts. The interesterification reaction was optimized with respect to the type and loading of biocatalyst. Equilibrium was approached in the shortest time with Novozym 435 (80% conversion in 4 h). The chemical, physical, and functional properties of the products were characterized. Appropriate choices of the reaction conditions and the initial ratio of EVOO to FHPO lead to TAG with melting profiles and solid fat contents similar to those of commercial products. Differences were observed in the solid fat contents, melting profiles, and oxidative stabilities of the various interesterified products and also between the indicated properties of each category of product and the corresponding physical blend of the precursor reagents.     

Aspergillus oryzae Lipase-Catalyzed Synthesis of Dioleoyl; Palmitoyl-Rich Triacylglycerols in Two Reactors

Dioleoyl; palmitoyl‐rich triacylglycerols (OPO‐rich TAG) were synthesized through Aspergillus oryzae lipase (AOL)‐catalyzed acidolysis of palm stearin with commercial oleic acid by a one‐step process in a stirred tank reactor and continuous packed bed reactor to evaluate the feasibility of using immobilized AOL. AOL was found to be valuable for the synthesis of OPO‐rich TAG when compared with commercial lipase from Thermomyces lanuginose (Lipozyme^® TL IM; Novozymes A/S, Bagsvaerd, Denmark). The C52 (triglycerides with a carbon number of 52, stands for OPO, OPL, LPL and their isomers) content of AOL was higher (45.65 %), and the intensity of treatment (IOT: lipase amount × reaction time/TAG amount) of AOL was just 6.25 % of that of Lipozyme^® TL IM under similar reaction conditions in the stirred tank reactor. Response surface methodology were used to optimize the reaction conditions of the AOL‐catalyzed acidolysis is reaction in the packed bed reactor. The optimal point for the set of experimental conditions generated were as follows: residence time 3.0 h; temperature 62.09 °C; substrate molar ratio 7.13 mol/mol. The highest C52 content obtained was 48.60 ± 2.36 %, with 57.46 ± 1.72 % total palmitic acid at the sn‐2 position and 74.21 ± 2.45 % oleic acid at the sn‐1,3 positions. The half‐life of AOL was 24 h in the stirred tank reactor and 140 h in the packed bed reactor. The immobilized AOL achieved similar conversion and selectivity to commercial lipases for the catalyzed synthesis of OPO‐rich TAG and may offer a cheaper alternative.     

Formulation and Characterization of Human Milk Fat Substitutes Made from Blends of Refined Palm Olein,and Soybean,Olive, Fish,and Virgin Coconut Oils

The simplest and the most cost-effective way of human milk fat substitute (HMFS) production is formulating of suitable vegetable oils at proper ratios. To do this, the D-optimal mixture design was used to optimize the HMFS formulation. The design included 25 formulations made from refined palm olein (35–55%), soybean oil (5–25%), olive oil (5–20%), virgin coconut oil (5–15%), and fish oil (0–10%). Samples were produced in laboratory and characterized in terms of fatty acid and triacylglycerol (TAG) compositions, free fatty acid content, peroxide value, iodine value, and oxidative stability index (OSI). HMFS samples were also compared with Codex Alimentarius (CA) and Iran National Standards Organization (INSO) standards. Each characteristic of HMFS samples was then expressed as a function of ingredient ratio using regression models. Finally, using numerical optimization, four optimized blends (PB₁-PB₄) were selected, made in the laboratory (HMFS₁-HMFS₄), characterized, and compared with CA and INSO standards. The properties of all the optimized blends (except the palmitic acid content of HMFS₂ and the monounsaturated fatty acid [MUFA] content of HMFS₃) met the standards. HMFS₄ showed the highest OSI in Rancimat and the lowest oxidation rate in Schaal oven tests. POL (19.53–21.73%), PPO (20.77–21.73%), OOO (9.11–11.16%), and OPO (8.84–9.46%) were the main (totally about 60%) TAG species found in HMFS samples. In conclusion, the HMFS₄ formula (55% palm olein, 13.5% soybean oil, 16% refined olive oil, 15% virgin coconut oil, and 0.5% fish oil) was suggested as the best formula for HMFS production.     

The influence of chemical interesterification on the physicochemical properties of complex fat systems. 3. Rheology and fractality of the crystal network

Chemically interesterified and noninteresterified lard-canola oil (LCO) and palm oil-soybean oil blends ranging from 100% hardstock to 50%:50% hardstock/vegetable oil (w/w) were evaluated for hardness index (HI) using cone penetrometry and viscoelastic properties, such as shear storage modulus G′, using controlled-stress rheometry. The HI and G′ of palm oil decreased upon addition of soybean oil, and chemical interesterification did not affect the HI or G′ of palm oil or palm oil-soybean oil blends. The HI and G′ of lard decreased upon addition of canola oil, while chemical interesterification led to increases in both the HI and G′ of lard and LCO blends. All these effects were nonsolid fat content-related, since solid fat content did not change substantially upon chemical interesterification. The microstructure of the fat crystal network in lard and palm oil was quantified rheologically using fractal analysis. Chemical interesterification did not affect the fractal dimension of the fat crystal networks in palm oil or lard (2.82 and 2.88, respectively). The rheological properties of the macroscopic systems were not affected by the spatial distribution of mass within their fat crystal networks. Moreover, our results suggest that the increases in G′ observed in lard upon chemical interesterification are potentially due to changes in the properties of the particles which make up the network (crystal habit).     

Oxidative stability of blends and interesterified blends of soybean oil and palm olein

Improvement of oxidative stability of soybean oil by blending with a more stable oil was investigated. Autoxidation of blends and interesterified blends (9∶1, 8∶2, 7∶3 and 1∶1, w/w) of soybean oil and palm olein was studied with respect to fatty acid composition, fatty acid location and triacylglycerol composition. Rates of formation of triacylglycerol hydroproxides, peroxide value and volatiles were evaluated. The fatty acid composition of soybean oil was changed by blending. Linolenic and linoleic acids decreased and oleic acid increased. The triacylglycerol composition of blends and interesterified blends was different from that of soybean oil. Relative to soybean oil, LnLL, LLL, LLO, LLP, LOO and LLS triacylglycerols were lowered and POO, POP and PLP were higher in blends and interesterified blends (where Ln, L, O, P and S represent linolenic, linoleic, oleic, palmitic and stearic acids, respectively). Interesterification of the blends leads to a decrease in POO and POP and an increase in LOP. Linoleic acid concentration at triacylglycerol carbon-2 was decreased by blending and interesterification. Rates of change for peroxide value and oxidation product formation confirmed the improvement of soybean oil stability by blending and interesterification. But, blends were more stable than interesterified blends. Also, the formation of hexanal, the major volatile of linoleate hydroperoxides of soybean oil, was decreased by blending and interesterification.     

Enzymatic Interesterification of Palm Oil and Fractions: Monitoring the Degree of Interesterification using Different Methods

Interesterification is an important modification technique for fats and oils resulting in the redistribution of the fatty acids among the glycerol backbone and thus changing the physico-chemical properties of the modified fat. In this study palm oil, palm olein and soft palm mid fraction (PMF) were subjected to both enzymatic (batchwise) and chemical interesterification. The reaction products were characterized before, during and after interesterification by HPLC, pulsed NMR (p-NMR) and differential scanning calorimetry (DSC). Interesterification led to more uniform triacylglycerol (TAG) compositions with smaller differences in final physico-chemical properties between the studied substrates. The degree of interesterification was evaluated on the basis of TAG composition and solid fat content (SFC). Significant differences between both methods were observed. The degree of interesterification based on SFC is therefore a better tool to evaluate the rate constant of the reaction as the TAG composition method does not take into consideration the formation of positional isomers at the end of the enzymatic process.     

Composition and Thermal Analysis of Lipids from Pre-fried Chicken Nuggets

Composition and thermal profiles of the endogenous lipids of ten commercial chicken nuggets brands (NPO, ACO, AFO, APO, ASO, AMO, ARO, JOD, SMO, and SOD) were compared with those of the lipids of chicken nuggets pre-fried in lard (ALD) and palm olein (AOO) to determine the type of oil used for pre-frying of the product. The stearic acid content of the commercial brands were similar to that of the sample pre-fried in palm olein, but significantly (p < 0.05) lower than that of the sample pre-fried in lard. The triacylglycerol (TAG) profiles of the commercial brands were similar to that of the sample pre-fried in palm olein, but distinctly different from the sample pre-fried in lard according to the dissimilarities in the contents of TAG molecules namely, PLL, POS, and PPO. Based on thermal analysis, the commercial brands of chicken nuggets could be divided into three distinguishable subgroups namely, Group-A: NPO; Group-B: ACO, AFO, APO, ASO; Group-C: AMO, ARO, JOD, SMO, SOD. While brands under group-B showed close similarity to AOO, none show any similarity to sample ALD. As any of the samples did not possess characteristics of the sample pre-fried in lard, the commercial brands of chicken nuggets of this study are recommendable for consumers whose religious restriction prevents the use of lard in food.     

TAG composition of ewe's milk fat. Detection of foreign fats

The TAG composition of 45 samples of ewe's milk, collected throughout the year from five Spanish breeds, was analyzed according to their carbon number by using short capillary column GC. The TAG content had a bimodal distribution with maxima at C₃₈ (12.8%) and C₅₂ (8.4%). The TAG composition did not vary significantly with respect to the time of year of sampling but was affected by the breed. Multiple regression equations based on TAG content are proposed to detect foreign fats in ewe's milk fat. Analysis of known mixtures of lard, palm oil, and cow's milk fat with ewe's milk fat have experimentally confirmed the accuracy of the equations.     

Performance of palm‐based fat blends with a low saturated fat content in puff pastry

Four fat blends based on palm fractions in combination with high oleic sunflower oil (HOSF) with a relatively low saturated fatty acid content (29.2 ± 0.85%, i.e. less than 50% of that of butter) were prepared. The saturated fat was located in different TAG structures in each blend. Principal saturated TAG were derived from palm stearin (POs, containing tripalmitoyl glycerol—PPP), palm mid‐fraction (PMF, containing 1,3‐dipalmitoyl‐2‐oleoyl glycerol—POP) and interesterified PMF (inPMF, containing PPP, POP and rac‐1,2‐dipalmitoyl‐3‐oleoyl glycerol—PPO). Thus, in blend 1, composed of POs and HOSF, the saturates resided principally in PPP. In blend 2, composed of POs, PMF and HOSF, the principal saturate‐containing TAG were PPP and POP. Blend 3, composed of inPMF and HOSF, was similar to blend 2 except that the disaturated TAG comprised a 2:1 mixture of PPO:POP. Finally, blend 4, a mixture of PMF and HOSF, had saturates present mainly as POP. The physical properties and the functionality of blends, as shortenings for puff pastry laminated in a warm bakery environment (20–24°C), were compared with each other, and with butter. Puff pastry prepared with blend 1 (POs:HOSF 29:71) and blend 4 (PMF:HOSF 41:59), was very hard; blend 2 (POs:PMF:HOSF 13:19:68) was most similar to butter in the compressibility of the baked product and it performed well in an independent baking trial; blend 3 (inPMF:HOSF 40:60) gave a product that required a higher force for compression than butter.     

TAG composition and solid fat content of palm oil, sunflower oil, and palm kernel olein belends before and after chemical interesterification

Modification of the characteristics of palm oil (PO), sunflower oil, and plam kernel olein (PKOo) according to conventional three-component mixture designs was undertaken by a combination of blending and chemical interesterification (CIE) techniques. TAG composition and solid fat content (SFC) profile of the starting blends were analyzed and compared with those of the interesterified blends. Upon CIE, extensive rearrangement of FA among TAG was evident. Concentrations of several TAG were increased, some were decreased, and several new TAG were formed. The resulting changes in TAG profile were reflected in the SFC of the blends. The SFC values of the chemically interesterified blends, except binary blends of PO/PKOo, revealed that they were softer than their respective starting blends. SFC data also indicated that eutectic interaction occurred between PO and PKOo in the starting blends and that this interaction was diminished after CIE.     

Human Milk Fat Substitute from Butterfat: Production by Enzymatic Interesterification and Evaluation of Oxidative Stability

Recent data have suggested that the fatty acid composition and molecular structure of fats in infant formulas should be as similar to human milk fat as possible to obtain optimal fat and calcium absorption from the infant formula. This work investigated the possibilities of using enzyme technology and butterfat as a material to produce a fat similar to human milk fat with respect to the above parameters. Moreover, the oxidative stability of the enzyme modified human milk fat substitute (HMFS) was compared to the fat blend used for the production of HMFS. Using a combination of enzyme technology, fractionation and batch deodorization and with butterfat in combination with soybean oil and rapeseed oil as raw materials it was possible to produce HMFS with a molecular structure and fatty acid composition that was very similar to that of human milk fat. The oxidative stability of the HMFS oil was lower than that of the reference oil with the same fatty acid composition. However, oxidation did not lead to a severe increase in rancidity scores during storage. Rather, the panel gave high intensity scores for other off-flavors such as burnt and bitter. Further optimization of the deodorization process is therefore necessary to remove these off-flavors.