Regiospecific Triacylglycerol Analysis: When Total Fatty Acid Composition Cannot Resolve the sn-2 Position

In lipid biochemistry, clinical nutrition, and specialty oleochemistry, relying solely on total fatty acid profiling produces dangerous analytical blind spots. Standard gas chromatography with flame ionization detection (GC-FID) of fatty acid methyl esters (FAMEs) measures only the gross percentage of fatty acids present in an oil. It provides zero information regarding stereospecific distribution across the glycerol backbone—whether a fatty acid resides at the outer primary positions (sn-1 and sn-3) or the internal secondary position (sn-2).

This stereochemical distinction governs physiological bioavailability and physical material behavior. Human pancreatic lipase is strictly sn-1,3 regiospecific; it selectively cleaves outer ester bonds, releasing two free fatty acids and one 2-monoacylglycerol (2-MAG). Saturated fatty acids (such as palmitic acid, 16:0) located at sn-1,3 form insoluble calcium soaps that cause severe infant constipation and nutrient loss, whereas sn-2 palmitate (as in human breast milk and structured OPO lipids) is absorbed with over 90% efficiency. Similarly, in single-cell algae oil, docosahexaenoic acid (DHA, 22:6n-3) esterified at sn-2 is absorbed as 2-DHA-MAG, providing preferential neurodevelopmental accretion across the blood-brain barrier. When total fatty acid composition cannot distinguish functional structured lipids from low-grade blends, regiospecific triacylglycerol analysis via high-resolution LC-MS/MS and non-aqueous chromatography provides the definitive solution.

Core Analytical Principles of Triacylglycerol Regiospecificity

A triacylglycerol (TAG) molecule consists of a glycerol backbone esterified with three fatty acid acyl chains. Because the central carbon atom of glycerol (C-2) possesses pro-chirality, IUPAC-IUBMB nomenclature enforces stereospecific numbering (sn):

• sn-1 Position: The primary hydroxyl group pointing to the top in a Fischer projection where the secondary hydroxyl at C-2 is oriented to the left.
• sn-2 Position: The internal secondary hydroxyl group at the central C-2 carbon.
• sn-3 Position: The primary hydroxyl group pointing to the bottom.

When a TAG contains two or three different fatty acyl chains (e.g., palmitic acid P and oleic acid O), multiple regioisomers exist:

Both POP and PPO share identical elemental formulas (C₅₃H₁₀₀O₆), identical exact molecular weights (832.75 Da), and identical total fatty acid percentages (66.7% palmitic, 33.3% oleic). Standard GC-FID or low-resolution mass spectrometry reports these two distinct chemical entities as a single unresolved sum species (TAG 50:1). However, their digestion kinetics, metabolic fates, and crystallization melting points differ fundamentally.

For research teams analyzing complex lipid fractions across diverse biological matrices, our dedicated untargeted lipidomics profiling service integrates high-resolution mass spectrometry to untangle complex lipidome architectures.

Figure 1. The digestive physiology of triacylglycerol regiospecificity in the human gastrointestinal tract. Left: Symmetrical OPO (human milk fat / structured lipid) undergoes sn-1,3 cleavage by pancreatic lipase to yield 2-mono-palmitoylglycerol, which is solubilized in mixed bile micelles and efficiently absorbed (>90%). Right: Asymmetrical PPO (standard vegetable oil) releases free palmitic acid at sn-1,3, which precipitates with ionized calcium at luminal pH (7.0–8.0) to form insoluble calcium soaps, resulting in fecal loss and constipation.

Digestive Bioavailability: The sn-2 Metabolic Advantage in Human Nutrition

The digestive processing of dietary fats in the mammalian intestine is governed by the structural specificity of digestive lipases:

Pancreatic Lipase Stereospecific Cleavage Kinetics

Gastric lipase in the stomach and pancreatic lipase in the duodenum operate with absolute sn-1,3 regioselectivity. The catalytic triad of pancreatic lipase (Ser152, Asp176, His263) can physically access only the sterically exposed primary ester bonds at the sn-1 and sn-3 positions of the glycerol backbone. It cannot hydrolyze the sterically shielded ester bond at sn-2.

Consequently, intestinal hydrolysis of one triacylglycerol molecule generates:

• Two Free Fatty Acids (FFAs) derived from the sn-1 and sn-3 positions.
• One 2-Monoacylglycerol (2-MAG) containing the original fatty acid from the sn-2 position intact.

The Mechanism of Insoluble Calcium Soap Formation

The physical fate of the released FFAs depends on their melting point and carbon chain saturation:

• Unsaturated FFAs (e.g., Oleic 18:1, Linoleic 18:2): Possess low melting points (<16°C) and form flexible liquid phases. At intestinal pH (~7.4), they readily incorporate into bile salt-phospholipid mixed micelles and diffuse across the enterocyte brush border membrane.
• Saturated FFAs (e.g., Palmitic 16:0, Stearic 18:0): Saturated long-chain fatty acids have high melting points (palmitic acid mp ~63°C; stearic acid mp ~70°C). When released as free acids at body temperature (37°C), they do not micellize easily. Instead, their ionized carboxylate headgroups (-COO^-) react immediately with divalent cations (Ca²⁺ and Mg²⁺) in the intestinal lumen to form insoluble fatty acid soaps (e.g., calcium dipalmitate, [CH₃(CH₂)₁₄COO]₂Ca).

Calcium soaps cannot be absorbed. They form hard, chalky precipitates in the infant bowel, leading to painful hard stools, chronic constipation, diminished bone mineral accretion due to calcium excretion, and reduced caloric recovery.

The sn-2 Palmitate and sn-2 DHA Biological Rescue

Human evolution solved this bioenergetic crisis through stereospecific lipid assembly. In human breast milk, over 70% of total palmitic acid is esterified at the sn-2 position (forming OPO and related structured lipids). When digested, palmitic acid remains esterified as 2-mono-palmitoylglycerol (2-MAG). Because 2-MAG possesses two free polar hydroxyl groups, it is highly amphiphilic, micellizes spontaneously without precipitating with calcium, and is absorbed across the enterocyte membrane with >90% efficiency. Once inside the enterocyte, 2-MAG serves as the direct backbone for re-esterification into chylomicrons via the monoacylglycerol pathway.

In neurodevelopment, dietary docosahexaenoic acid (DHA, 22:6n-3) esterified at sn-2 in maternal milk and marine single-cell algae oil is absorbed as 2-DHA-MAG. After enterocyte re-synthesis, a significant fraction is channeled into lysophosphatidylcholine-DHA (LPC-DHA), which is the specific physiological ligand recognized by the Mfsd2a transporter at the blood-brain barrier. DHA supplied at the sn-2 position exhibits up to 2- to 3-fold higher brain and retinal accretion compared to random ethyl esters or sn-1,3 DHA.

Analytical Boundaries: Total Fatty Acid Profiles (GC-FID) vs. Positional TAGs

Standard quality control in food science and industrial oleochemistry relies on gas chromatography of fatty acid methyl esters (AOCS Official Method Ce 1h-05 / AOAC 996.06). While fast and reproducible for bulk commodities, total FAME analysis is inherently blind to regiospecificity.

Why Saponification Erases Stereochemical Information

To prepare samples for GC-FID, the triacylglycerol oil is subjected to alkaline saponification (e.g., KOH in methanol) or acid-catalyzed transesterification (BF₃ or H₂SO₄ in methanol). This reaction hydrolyzes all three ester bonds indiscriminately, pooling fatty acids from sn-1, sn-2, and sn-3 into a single homogeneous pool of methyl esters.

Consider two infant formula lipid blends with identical total FAME profiles: 35% Palmitic (16:0), 45% Oleic (18:1), and 15% Linoleic (18:2).

• Blend A (Premium Structured OPO Fat): 72% of total palmitic acid is esterified at sn-2. In clinical trials, infants fed Blend A exhibit soft stools, reduced crying, and superior calcium absorption matching breastfed infants.
• Blend B (Cheap Palm Olein / Soy Blend): Only 14% of palmitic acid is at sn-2 (>86% at sn-1,3). In clinical trials, infants suffer high rates of constipation and calcium soap excretion.

Under standard GC-FID analysis, Blend A and Blend B produce identical chromatograms. Total fatty acid analysis cannot detect the fraudulent substitution of cheap vegetable oils for expensive structured lipids, making regiospecific analysis mandatory for regulatory compliance and product release.

For detailed lipid class quantification and targeted profiling, explore our specialized phospholipids analysis service, providing comprehensive molecular species resolution.

Figure 2. Gas-phase kinetic fragmentation of triacylglycerol ammonium adducts ([M+NH4]+) in tandem mass spectrometry. Left: Symmetrical POP undergoes preferential neutral loss of palmitic acid from sn-1 or sn-3, generating an intense diacylglycerol fragment ion [DAG-PO]+ at m/z 577.5, with minimal loss of sn-2 oleic acid. Right: Asymmetrical PPO displays prominent loss of sn-1,3 oleic acid, yielding an intense [DAG-PP]+ fragment at m/z 551.5. Calibrating the fragment intensity ratio against synthetic regioisomers delivers direct sn-2 quantification without chemical digestion.

Enzymatic Hydrolysis Methods: Pancreatic Lipase Assays & Acyl Migration Risks

For decades, the standard method for determining the sn-2 position has been enzymatic hydrolysis using porcine pancreatic lipase, codified in AOCS Official Method Ch 3-91. While historically groundbreaking, this classical wet-chemistry assay suffers from severe physical limitations:

1. Spontaneous Acyl Migration Artifacts (2-to-1 Shifting)

The enzymatic hydrolysis of TAG yields 1,2-diacylglycerols (1,2-DAG), which are subsequently cleaved to 2-monoacylglycerols (2-MAG). The 2-MAG intermediate possesses a free secondary ester. Thermodynamically, a 2-monoester is less stable than a 1-monoester due to steric crowding and intramolecular hydrogen bonding:

During the pancreatic lipase incubation (typically 5 to 15 minutes at 40°C), liquid-liquid extraction, and silica-gel thin-layer chromatography (TLC) band recovery, between 10% and 25% of the fatty acids at the sn-2 position spontaneously migrate to sn-1(3). This migration leads to severe underestimation of true sn-2 occupancy. If the assay is prolonged or the pH drifts above 8.0, the apparent sn-2 palmitate content can drop by half.

2. Severe Steric Hindrance Against Long-Chain PUFAs (DHA & EPA)

Porcine pancreatic lipase exhibits extreme substrate specificity. Its active site pocket readily accepts straight-chain saturated fatty acids (14:0, 16:0, 18:0) and monounsaturated fatty acids (18:1). However, long-chain polyunsaturated fatty acids—particularly DHA (22:6n-3) and EPA (20:5n-3)—possess bulky, curved conformations with multiple methylene-interrupted double bonds.

When DHA is present at the sn-1 or sn-3 positions, pancreatic lipase hydrolyzes the ester bond at a rate 10- to 50-fold slower than palmitate. Consequently, in marine fish oils and Schizochytrium algae oils, standard pancreatic lipase digestion fails to reach completion. Unhydrolyzed diacylglycerols and intact TAGs contaminate the 2-MAG fraction, yielding completely erroneous sn-2 PUFA values.

Tandem Mass Spectrometry: Gas-Phase Cleavage Kinetics of [M+NH4]+ Adducts

Modern regiospecific lipidomics replaces wet-chemistry enzymatic digestion with high-resolution tandem mass spectrometry (LC-MS/MS). By interrogating intact, unhydrolyzed triacylglycerol ions in the gas phase, LC-MS/MS eliminates acyl migration completely and accurately measures long-chain PUFAs.

Electrospray Ionization and Ammonium Adduct Formation

Intact TAGs lack basic functional groups and do not form stable protonated molecules ([M+H]⁺) efficiently in standard electrospray ionization (ESI), often undergoing extensive in-source fragmentation. By supplementing mobile phases with 5 to 10 mM ammonium formate or ammonium acetate, TAGs ionize predominantly as intact ammonium adducts ([M+NH₄]⁺).

Kinetic Preference for Outer sn-1,3 Fatty Acid Neutral Loss

When an [M+NH₄]⁺ ion is accelerated into a collision cell (collision-induced dissociation, CID), the initial energetic step is the loss of neutral ammonia (NH₃) to form a transient protonated species [M+H]⁺. Subsequent collision activation drives the elimination of one of the three fatty acid acyl chains as a neutral free fatty acid (R_iCOOH), generating a pair of diacylglycerol-like fragment ions ([DAG]⁺):

Crucially, gas-phase fragmentation kinetics follow strict stereochemical energy profiles:

• Outer sn-1 and sn-3 Ester Cleavage: The primary ester bonds at the outer ends of the glycerol backbone are exposed to minimal steric hindrance. Their cleavage requires lower activation energy ($E_{\text{a}}$), making neutral loss from sn-1 and sn-3 kinetically favored.
• Internal sn-2 Ester Cleavage: The secondary ester bond at sn-2 is sterically shielded by the adjacent acyl chains. Cleavage requires higher activation energy, proceeding at a significantly lower reaction rate.

Regioisomeric Deconvolution: Symmetrical vs. Asymmetrical Pairs

Consider the regioisomeric pair POP (1,3-dipalmitoyl-2-oleoylglycerol) and PPO (1,2-dipalmitoyl-3-oleoylglycerol) ([M+NH₄]⁺ at $m/z = 850.78$):

• In POP (symmetrical, Oleic at sn-2): Loss of palmitate from sn-1 or sn-3 produces the diacylglycerol fragment $[\text{PO}]^+$ at $m/z = 577.5$. Loss of oleate from the protected sn-2 position would produce $[\text{PP}]^+$ at $m/z = 551.5$. In experimental CID spectra of pure POP, the $[\text{PO}]^+$ fragment dominates, while $[\text{PP}]^+$ is near-baseline (<10% relative intensity).
• In PPO (asymmetrical, Palmitic at sn-2): Oleic acid resides at the outer sn-3 position. Because outer loss is kinetically favored, neutral loss of oleate occurs readily, producing a prominent $[\text{PP}]^+$ fragment at $m/z = 551.5$.

By measuring the intensity ratio $I([\text{PP}]^+) / I([\text{PO}]^+)$ and plotting against synthetic standard calibration curves, the fractional proportion of PPO vs POP in an unknown oil sample is quantified with high precision (±1.5%), completely bypassing enzymatic hydrolysis.

Chromatographic Isomer Resolution: Silver-Ion (Ag+) HPLC, NARP-LC, and SFC

While mass spectral fragmentation ratios resolve many regioisomers, chromatographically separating intact TAG isomers prior to mass spectrometry provides the highest analytical confidence:

1. Non-Aqueous Reversed-Phase (NARP) UHPLC

Standard aqueous mobile phases cannot dissolve hydrophobic triacylglycerols. NARP chromatography utilizes 100% organic mobile phases—typically gradients of Acetonitrile / Methanol / Isopropanol—combined with high-density octadecyl (C18) or triacontyl (C30) stationary phases. C30 phases feature extended alkyl chains that embed deeply into the hydrophobic clefts of triacylglycerols, achieving partial to complete chromatographic resolution of symmetrical and asymmetrical pairs (e.g., separating POP from PPO) based on molecular shape and curvature.

2. Silver-Ion High-Performance Liquid Chromatography (Àg⁺-HPLC)

In silver-ion chromatography, stationary-phase sulfonic acid groups are loaded with silver ions (Ag⁺). The Ag⁺ cations form reversible, non-covalent π-complexes with the double bonds of unsaturated fatty acyl chains. The retention strength is governed by:

• The total number of double bonds in the TAG molecule.
• The geometric configuration (cis double bonds form stronger π-complexes than trans double bonds).
• The steric accessibility of the double bond (a double bond at the outer sn-1,3 positions interacts more strongly with the stationary phase than one buried at sn-2).

Consequently, Ag⁺-HPLC provides baseline resolution of complex mixtures that are co-eluting on standard reversed-phase columns, such as separating OPO from OOP, or resolving trans-fatty acid isomers.

3. Supercritical Fluid Chromatography (SFC)

Supercritical CO₂ combined with methanol/isopropanol co-solvents provides high diffusivity and low viscosity. Modern SFC systems coupled to triple quadrupole or Q-TOF mass spectrometers achieve baseline separation of triacylglycerol positional regioisomers in under 8 minutes, making it the premier platform for high-throughput industrial release testing.

Figure 3. Analytical workflow for regiospecific TAG analysis in single-cell algae oil (Schizochytrium sp.). Top: Non-aqueous reversed-phase UHPLC separation of DHA-containing triacylglycerol species (DHA-DHA-DHA, DHA-DHA-16:0, DHA-DHA-EPA). Bottom: Tandem mass spectrometric deconvolution using diacylglycerol fragment ratios to verify sn-2 DHA positional occupancy, supporting infant formula raw material qualification and cognitive health claims.

Industrial Quality Control: Algae Oil (sn-2 DHA), OPO Infant Formula & CBE Fats

Regiospecific triacylglycerol analysis is an active regulatory and quality-assurance gate across multi-billion-dollar global markets:

1. Single-Cell Algae Oil (Schizochytrium sp. DHA)

Marine microalgae, particularly Schizochytrium sp. and Crypthecodinium cohnii, are industrially cultivated to produce single-cell oils containing 40% to 55% DHA by weight. These oils are the primary vegetarian source of omega-3 long-chain PUFAs for infant formulas and prenatal supplements.

Total fatty acid GC-FID confirms that an oil contains 40% DHA, but cannot tell whether the DHA is located at sn-2. For infant cognitive and retinal development, formula manufacturers mandate that at least 45% to 50% of the DHA reside at the sn-2 position. Testing laboratories deploy NARP-LC-MS/MS to quantify individual molecular species—such as DHA/DHA/DHA (tridocosahexaenoin), DHA/DHA/16:0, and DHA/DHA/14:0—and confirm sn-2 DHA compliance.

2. Structured Human Milk Fat Substitutes (OPO / SN-2 Palmitate)

In standard bovine milk and cheap palm oil, palmitic acid is predominantly esterified at the outer sn-1 and sn-3 positions (<20% at sn-2). To prevent infant constipation and calcium deficiency, specialty oleochemical manufacturers produce structured OPO fat (1,3-dioleoyl-2-palmitoylglycerol) via 1,3-specific enzymatic interesterification of tripalmitin with oleic acid using immobilized lipases (e.g., Rhizomucor miehei lipase).

To qualify as commercial infant formula grade OPO fat (e.g., Infat®, Betapol®), the product must pass stringent release testing demonstrating that ≥65% to 70% of total palmitic acid is positioned at sn-2, with symmetrical OPO exceeding asymmetrical PPO by a factor of 4 to 1.

3. Cocoa Butter Equivalents (CBE) and Confectionery Fats

Natural cocoa butter possesses a unique chemical composition consisting of over 80% symmetrical monounsaturated triacylglycerols: POP (1,3-dipalmitoyl-2-oleoylglycerol), POS, and SOS. This strict symmetrical architecture allows the fat to crystallize into the polymorphic Form V (β₂) crystal lattice, which melts sharply at body temperature (34–36°C) to provide chocolate's characteristic snap and glossy texture.

When formulating Cocoa Butter Equivalents (CBE) from palm mid-fractions and shea stearin, the presence of asymmetric isomers (PPO and SSO) disrupts the crystal lattice, causing product softening, delayed demolding, and severe fat bloom (surface recrystallization). Quality control laboratories mandate LC-MS/MS testing to verify that asymmetrical PPO content remains strictly below 2.0%.

Quantitative Deconvolution Mathematics: sn-2 Positional Occupancy Formulas

To translate tandem mass spectrometry data into reportable percentages of sn-2 occupancy, analytical laboratories utilize calibrated mathematical deconvolution models:

1. Symmetrical vs. Asymmetrical Regioisomeric Ratio

For a binary mixture of regioisomers containing fatty acids A and B (e.g., ABA vs AAB, such as POP vs PPO), fragmentation generates two diagnostic diacylglycerol ions:

• $[\text{DAG}_{\text{AB}}]^+ = [\text{M} + \text{H} - \text{A}]^+$ (Loss of outer acid A from ABA, or outer/inner from AAB).
• $[\text{DAG}_{\text{AA}}]^+ = [\text{M} + \text{H} - \text{B}]^+$ (Loss of outer acid B from AAB).

The relative fragment intensity ratio ($R$) is defined as:

Using a calibration curve established with pure synthetic standards (e.g., pure sn-POP mixed with pure sn-PPO in ratios from 0:100 to 100:0):

2. Calculating Relative sn-2 Fatty Acid Distribution

Across an entire lipid sample, the percentage of a specific fatty acid $i$ esterified at the sn-2 position relative to total fatty acid $i$ is calculated by integrating individual TAG species abundances:

This metric directly populates commercial Certificates of Analysis (CoA) for infant formula structured lipids and algae oils.

Technical Troubleshooting: Incomplete Digestion, Acyl Shifting & Ion Suppression

Navigating common failure modes in regiospecific analysis ensures defensible analytical reports:

1. Identifying and Correcting Acyl Migration in Enzymatic Assays

If utilizing the traditional pancreatic lipase method:

• Verify TLC Purity: Run authentic 1,2-DAG, 1,3-DAG, 2-MAG, and 1-MAG reference standards on every silica gel TLC plate. Stain with 0.05% rhodamine 6G to ensure baseline separation of the 2-MAG band.
• Control Incubation Time: Stop lipase digestion strictly at 60% to 70% total hydrolysis (typically 3 to 5 minutes at 40°C). Over-digesting beyond 8 minutes exponentially accelerates 2-to-1 acyl shifting.
• Acidify During Extraction: Immediately quench the enzymatic reaction by adding 1 mL of 1 M HCl. Acyl migration is kinetically minimized at acidic pH (3.0–4.0).

2. Managing Ionization Suppression in LC-MS/MS

Triacylglycerols are non-polar and elute late in chromatographic runs where mobile phase organic content exceeds 90%. If sample extracts contain high concentrations of free sterols, phospholipids, or residual solvent impurities:

• Perform pre-cleanup using a solid-phase extraction (SPE) silica cartridge to elute neutral TAGs with hexane/diethyl ether (95:5 v/v), discarding polar phospholipids and free fatty acids.
• Ensure continuous post-column addition of 5 mM ammonium formate to maintain constant ionization efficiency throughout the hydrophobic gradient.

Sample Submission & Method Selection Guidelines

• Sample Volume Requirements:
  • Pure Bulk Oils (Refined Algae Oil, Structured OPO, CBE Fats): 1.0 to 2.0 mL.
  • Powdered Infant Formulas: 50 to 100 g (requires gentle Folch/Mojonnier extraction prior to analysis).
  • Freeze-Dried Biomass (Schizochytrium microalgae paste): 5 to 10 g.
• Protect Polyunsaturated Fatty Acids from Peroxidation:
  • Add 0.01% (w/v) butylated hydroxytoluene (BHT) to all extraction solvents.
  • Purge sample headspace with inert nitrogen or argon gas prior to sealing vials.
  • Store samples at -80°C and ship on dry ice. Avoid repeated freeze-thaw cycles.
• Method Selection Matrix:
  • For Regulatory Infant Formula Release (OPO/sn-2 Palmitate): Select NARP-UHPLC-MS/MS Direct Regioisomeric Analysis for rapid, migration-free certification.
  • For High-PUFA Algae Oils (DHA/EPA): Select Direct Intact LC-MS/MS with C30 Separation; pancreatic lipase is strictly contraindicated due to PUFA resistance.
  • For Chocolate Confectionery Fats (CBE/POP/POS/SOS): Select Silver-Ion HPLC with Evaporative Light Scattering Detection (ELSD) or Mass Spectrometry.

References:

1. Mu, H., & Høy, C. E. (2004). The digestion of dietary triacylglycerols. Progress in Lipid Research, 43(2), 105–133. doi:10.1016/S0163-7827(03)00050-X.
2. Innis, S. M. (2011). Dietary triacylglycerol structure and its role in infant nutrition. Advances in Nutrition, 2(3), 275–283. doi:10.3945/an.111.000448.
3. Leskinen, H., et al. (2007). Regiospecific analysis of triacylglycerols using silver ion high-performance liquid chromatography and atmospheric pressure chemical ionization mass spectrometry. Rapid Communications in Mass Spectrometry, 21(14), 2361–2373. doi:10.1002/rcm.3097.
4. Nguyen, V. C., et al. (2014). Quantification of triacylglycerol molecular species in infant formula by liquid chromatography-tandem mass spectrometry. Journal of Agricultural and Food Chemistry, 62(41), 10125–10134. doi:10.1021/jf503522u.
5. Mawatari, S., et al. (2020). Positional distribution of DHA in triacylglycerols from single cell oils and marine lipids. Lipids, 55(4), 381–392. doi:10.1002/lipd.12245.
6. American Oil Chemists' Society (AOCS). (2017). Determination of fatty acids in the 2-position of triacylglycerols (Official Method Ch 3-91). Official Methods and Recommended Practices of the AOCS, 7th Edition. AOCS Standards.
7. Linderborg, K. M., et al. (2019). Regiospecific analysis of triacylglycerols: A review of methodology and dietary implications. Trends in Food Science & Technology, 88, 521–534. doi:10.1016/j.tifs.2019.04.015.