Comparative analysis of plant and animal milks: evidence of intermolecular interactions

Comparative analysis of plant and animal milks: evidence of intermolecular interactions | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL Sustainable Food Proteins This is a preprint and has not been peer reviewed. Data may be preliminary. 29 January 2025 V1 Latest version Share on Comparative analysis of plant and animal milks: evidence of intermolecular interactions Authors : Tina Kayeye 0009-0003-3860-4918 [email protected] , Alison Rodger 0000-0002-7111-3024 , Yuling Wang 0000-0003-3627-7397 , Anwar Sunna , John Ashton , Andrew Penton , and Sarah Noyes Authors Info & Affiliations https://doi.org/10.22541/au.173816988.81450559/v1 580 views 265 downloads Contents Abstract Introduction Quality control of plant-based milks Nutritional information panels on products Intermolecular structures in bovine milk Total milk analysis Separation plant-based milk components Centrifugation and filtration Chemical extraction Chromatographic separation Content analysis of plant-based milks Conclusion and Future Perspectives Acknowledgement Supplementary Material References Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract Commercially available plant-based milks are milk analogues manufactured to mimic the emulsion structural, nutritional, and functional properties of animal milks by combining plant materials extracted from e.g. oat or almond combined with added fats, vitamins and minerals. It is not yet clear how the molecules in plant-based milks interact to form the milk structures or even what the composition of the final products is. The purpose of this review is to bring together evidence relating to the molecular composition of plant-based milks and the structural arrangement of the molecules in the final products. Much of the available work involves chromatographic methods, microscopy, gravimetric, volumetric and elemental analyses methods. Although confocal microscopy with fluorescence labelling has been used to stain for different molecules in milk, there is little information on inter-molecular interactions. There is thus a need for new methods to investigate component interactions to ensure the maintenance of plant milk product quality. Introduction Plant-based milks, such as oat, soy and almond, are fabricated emulsion systems produced to mimic the appearance, behaviour, and sensory profile of animal milks (McClements, 2019). They are water extracts of plant material (McClements & Grossmann, 2021) with added vegetable oils such as canola and sunflower, buffers to neutralise acidity, and added minerals (such as calcium and phosphorus) to improve their nutritional value. Many people are gravitating to plant-based milk consumption due to lifestyles change (vegans and vegetarians), or concern about high levels of saturated fats (MacGibbon & Taylor, 2009) and cholesterol (Artaud-Wild, Connor, Sexton, & Connor, 1993), or concern about the environmental issues associated with animal milk production (Poore & Nemecek, 2018; Rotz, Montes, & Chianese, 2010) ( Fig. 1 ). Furthermore, some individuals are unable to break down lactose due to a lack of the enzyme required to digest this sugar, which is the most abundant in animal milks. As a result, they may experience digestive discomfort after consuming animal milks and are increasingly seeking alternatives (Bodé & Gudmand-Høyer, 1988; Jansson-Knodell, Krajicek, Savaiano, & Shin, 2020; Swagerty, D.Walling, & Klein, 2002). For plant-based milk products to be widely accepted, they must meet the consumer’s expectations on taste, mouthfeel and function properties. Therefore, plant milk manufacturers often strive to produce a milk that is comparable to animal milk. This involves the addition of vegetable oils, minerals and gums to improve the taste and appearance of the milk, as well as homogenisation and ultra-high temperature treatment and finally aseptic packaging. Plant milks are therefore made up of the building blocks proteins, fats and carbohydrates which are found in animal milks, but they may not be combined in the same way in plant-based milks. Molecular associations contribute to the characteristic flavours, tastes, mouthfeel, stability, and milk product behaviour during food applications. Consumers expect plant based milks to behave similarly to cow’s milk, especially when preparing cappuccino-style drinks (Jaeger, Dupas de Matos, Frempomaa Oduro, & Hort, 2024). Researchers have reported that consumers find commercially available coconut/soy blend milk and oat milk to have inferior foaming properties compared to soy and bovine milk at 65°C, the temperature typically used for coffee/milk drinks (Zakidou, Varka, & Paraskevopoulou, 2022). Some plant-based milks even separate when poured into coffee, presumably due to its acidity. For some products, simple pH buffering works well but for others there is a complicated interdependency between emulsifiers present and formulation stability. For example, oat milks tend to have poor emulsion stability mainly because of the high starch content of oats (Deswal, Deora, & Mishra, 2014; Sethi, Tyagi, & Anurag, 2016). Also, oats are low in calcium, therefore manufacturers add minerals to improve their nutritional value (Sethi et al., 2016; Tangyu, Muller, Bolten, & Wittmann, 2019). However, little is known on how these minerals interact with the building blocks in the plant milks products, so their bioavailability is unknown. To our knowledge, no existing review article brings together evidence of molecular content and intermolecular associations in animal and plant-based milks, with a particular focus on the analytical methods used for their study. Accordingly, this review aims to compile evidence from the literature on the molecular composition of plant-based milks and the structural arrangement of their molecules in the final products, emphasising the techniques used for molecular component analysis. Special attention is given to analytical methods for assessing milk molecular components and total milk quality control (QC). Quality control of plant-based milks To meet regulatory requirements and ensure consumer satisfaction, plant milk manufacturers perform quality control (QC) measures of their products before releasing them to the market. Typically, plant-based milk product characterisation and QC involve the analysis of milks without separation of molecular species, but with monitoring of physical parameters such as appearance, particle size, viscosity as well as sensory profile. Various analysis methods are used. For example, much of the QC testing of oat milk involves controlling enzymatic digestion time, ingredients used and processing conditions. Typical milk QC tests include gravimetric and volumetric tests combined with chemical analyses for total protein, fat and carbohydrate content, specific mineral content (e.g. calcium, phosphorus, sodium and potassium), dietary fibre and sugar content. Also, manufacturers report the saturated fat, unsaturated fat and lactose content (for lactose-intolerant consumers) on the nutritional information panel (NIP). Grossman et al. have written an excellent review on standardised quality control methods for testing plant-based milks, which covers a range of spectroscopic, microscopic and colorimetric techniques (Grossmann, Kinchla, Nolden, & McClements, 2021). Sample preparation for milk analysis generally involves chemical extraction and/or combustion to attain total values. The focus of the assays is typically on total content of the main components and usually involves the destruction of intermolecular structures and interactions within the milks. Plants naturally have a plethora of complex intermolecular interactions and small concentrations of a wide range of molecules such as polyphenolic compounds which may interact with proteins (Tosif et al., 2021) both covalently and non-covalently (Zhang, Huang, Wang, Wan, & Wang, 2024). The process of plant milk production involves heating at high temperatures, agitation, mixing, and pH adjustment, all of which can alter the chemical bonds and associations between plant molecular species. These changes may cause protein unfolding and the exposure of amino acids to other milk components, either during processing or storage (Zhang et al., 2024). The way molecules interact, and combine are difficult to measure but play a crucial role in the taste, flavour and stability of plant-based milk products. As a result, the industry relies on trained sensory panels to ensure product quality is maintained. However, this ‘analytical method’ can be biased and is costly when scaled up. Nutritional information panels on products Table 1 shows various ultra-high temperature treated (UHT) plant-based milk products found in Australian supermarkets with their major macromolecular content (protein, fat and carbohydrates) and minerals as shown on label claim. Although plant-based milk manufacturers verify raw material suppliers for oats, almonds and soybeans through audits, there are inevitable variations in the raw material due to external factors such as environmental conditions (e.g., global warming and changes in weather patterns), geographical location, and the amount of rainfall received for crop grow. Also, pre-processing treatment such as de-hulling and milling, which involve high heat or chemical treatments, can affect the molecular composition of plant materials, ultimately influencing the quality of the finished plant milk product. Depending on the plant material used for plant milk manufacturing and formulation, the macromolecular and mineral content of these products may differ. For example, oats have a high starch content (Tester & Karkalas, 1996) and can cause gelatinisation problems at high temperatures during processing. Starch is composed of long-chain polysaccharides amylose and amylopectin which are therefore enzymatically cleaved during oat milk production. Enzymes such as β-amylase are used to cleave the polysaccharides into simple sugars ( Fig. 2 ), which can be measured and quantified. The information provided by manufacturers on the long-life milk NIPs is guided by laws and regulations specific to each country. Australian milks, for example, may or may not have information on phosphorus or calcium content (Table 1). As there is variation in the macromolecule content of various plant milk products on the Australian market due to the different plant sources used in their production, to quote plant-based milk component content on the NIP, approved QC analytical methods must be used. Analytical QC methods involve the chemical and mechanical separation of milk components to identify and quantify each component separately, ensuring compliance with regulatory obligations. Table 1 Protein, fat, carbohydrate and mineral content (phosphorus, sodium and calcium) of some Australian market commercially available plant-based milk products. So Good Creamy Oat Sanitarium Health Food Company 0.7 2.6 6.5 100 45 120 The Alternative Dairy Co. Barista Soy Sanitarium Health Food Company 3.2 3.0 4.3 100 47 57 The Alternative Dairy Co. Barista Almond Sanitarium Health Food Company 2.8 2.6 2.8 100 38 94 The Alternative Dairy Co. oat milk Sanitarium Health Food Company 0.8 3.0 6.1 100 48 120 Milklab Almond Noumi Limited 0.8 2.5 2.8 ND 40 120 Vitasoy almond milk Vitasoy Australia Products 0.7 2.2 0.3 ND 43 120 Vitasoy UHT oat milk (unsweetened) Vitasoy Australia Products 3.2 3.0 5.8 110 60 160 Bonsoy Longlife Soy Spiral Foods (Japan) 4.1 2.2 5.5 ND 47 25 So Good Oat (high in calcium) Sanitarium Health Food Company 0.8 1.9 6.3 100 45 120 Australia’s Own Coconut Noumi Limited 0.1 1.8 0.2 ND 44 ND PureHarvest Organic Soy unsweetened PureHarvest 3.0 3.4 1.5 ND 15 120 PlantWell Soy (bone & gut) Sanitarium Health Food Company 4.1 3.2 4.1 137 51 200 Chobani Oat Chobani Pty Ltd 0.8 3.4 7.6 87 53 120 Califia Farms Oat Barista blend Califia farms (Spain) 0.7 3.0 5.7 ND 0.12 104 Oatly oatmilk (Sweden) 1.1 2.8 7.0 ND 40 ND *ND means the amount is not quoted on the NIP. Intermolecular structures in bovine milk Animal milks such as bovine milk have been extensively characterised, especially their lipidic structures (milk fat globules: MFGs) and casein micelles. These structures scatter light, giving the characteristic white appearance of bovine milk which is desired by many (McCarthy & Singh, 2009). In addition, the size of the particles plays a key role in the milk ‘feel’ and behaviour during consumption. Over the years, the structure of animal milk MFGs have been studied and a circular structure with triglycerides enclosed in a membrane has been proposed as shown in Fig. 3A . Researchers have proposed two mechanisms for their formation. The first mechanism involves the budding of triglyceride bodies enclosed in a monolayer of phospholipids surrounded by an apical membrane (Bargmann & Knoop, 1959; I. H. Mather & Keenan, 1998). The second is a secretory vesicles mechanism, where vesicles containing both caseins and lipid globules are expelled from the epithelial cell through exocytosis (F. B. P. Wooding, 1971, 1973). In both mechanisms, the secreted oil bodies end up enclosed in a plasma membrane (milk fat globule membrane, MFGM) from the originating cell and the caseins end up as casein particles assemblies. To understand these structures, researchers have isolated the MFGM in animal milks using high or low salt buffers, detergents (Freudenstein et al., 1979), enzymatic breakdown of proteins (I. H. Mather & Keenan, 1998) and/or organic solvents (F. Wooding & Kemp, 1975). In all these cases, the presence of proteins on the membrane of animal milks has been confirmed using techniques such as sodium dodecyl polyacrylamide gel electrophoresis (SDS-PAGE) of MFGM proteins, gas-liquid chromatography of proteins and sugars, and electron microscopy (Freudenstein et al., 1979). The presence of the transmembrane glycoprotein butyrophilin (BTN) in bovine MFGs has been identified with immunoperoxidase localisation assays using guinea pig antibodies (purified IgG) to BTN combined with electron microscopy (Franke et al., 1981). Other membrane proteins such as xanthine oxidase/dehydrogenase (XDH/XO), periodic acid Schiff 6/7 (PAS 6/7) glycoproteins, and fatty acid-binding proteins (FABP) have been found to be devoid of membrane-spanning domains using techniques such as lactoperoxidase radioactivity labelling of PAS 6/7 (Ian H. Mather & Keenan, 1975) and xanthine oxidase activity on different MFGM fractions (Ian H. Mather, Weber, & Keenan, 1977). Based on their origin, biochemical pathway and tests performed, researchers have used these insights to depict the structure of MFGs as shown in Fig. 3A . Plant-based milks emulsions are, by way of contrast, not intrinsic to the plant but created during the milk production process. As a result, they are oil-in-water emulsions containing added vegetable oils. Since they are plant-based, they also contain intrinsic oil bodies (oleosomes). These lipidic structures found naturally in plant-based milks are shown in Fig. 3B . Researchers have identified oil bodies in plant-based milks and characterised their shape, dispersion and edge smoothness. As global consumption of plant-based milks continues to rise, researchers are focusing on improving these products’ emulsion systems, meeting regulatory obligations, enhancing customer appeal, and developing analytical methods for final product QC. Total milk analysis As discussed above, plant milks are fabricated emulsions of fat, proteins and carbohydrates in in water, with added vitamins and mineral. The physicochemical parameters of these emulsions, such as particle size, shape and surface charge are crucial for the product’s mouthfeel and overall stability. Therefore, manufacturers usually aim to achieve particles sizes under 1 µm after homogenisation (Fellows, 2017), similar to bovine milk, as larger particles would result in a “gritty” sensation in the mouth (Grossmann et al., 2021). Techniques for measuring particle size and surface charge are discussed below. It is clear that when microscopy techniques, such as Confocal Laser Scanning Microscopy (CLSM), are used, no prior chemical separation of the milks is required. However, for milk component analysis, these molecular species must be extracted using buffers, organic solvents and acids. pH of plant milks Plant milks are water-extracted products primarily composed of water. As a result, their pH may be acidic, depending on the water used in production. To adjust the milk to a normal range, minerals, gums and phosphate buffers are added. Buffers are typically used to neutralise any acidity, extend shelf life, and enhance stability when the milks are exposed to acidic, high-heat conditions, such as in coffee, tea, or cooking. Plant milks primarily contain organic compounds with a backbone of carbon, nitrogen, oxygen and hydrogen atoms. During the QC for market release or research, the physicochemical properties, such as pH, are measured using a pH meter. These milks typically have a pH of about 7.0 (Dhakal, Giusti, & Balasubramaniam, 2016; Ho, Bhandari, & Bansal, 2023).When the milks become contaminated by microorganisms or chemically degraded, measuring pH serves as a simple technique to assess product quality. Particle size: light scattering Dynamic light scattering (DLS) and static light laser diffraction (LD) are by far the most used light scattering techniques for measuring the particle size distributions of plant-based milks. DLS is an intensity-based method that determines the hydrodynamic size of particles by analysing their motion in solution, using the Stokes-Einstein equation (Shnoudeh et al., 2019). Although DLS is non-invasive, it is a low-resolution technique due to its bulk intensity measurements. Similar to DLS, LD is a particle sizing technique. However, it determines particle size based on the angle of scattered light, with smaller particles producing smaller scattering angles compared to larger particles. According to the Mie theory, the refractive indices (both real and imaginary) of the particles must be known, enabling LD to achieve higher resolution for specific particle types, such as milk fat globules. LD is the preferred technique for analysing plant-based milks, particularly for research and QC. Unlike DLS, LD provides volume-based results, reporting particle size distributions as cumulative percentiles (e.g., D 10 , D 25 , D 50 , D 75 , D 90 ), which indicate which the proportion of particles smaller than or equal to a given size. Since particle size directly impacts both mouthfeel and product stability, numerous studies have utilised DLS and LD to characterise particle sizes (Kupikowska-Stobba, Domagała, & Kasprzak, 2024) (Devnani, Ong, Kentish, & Gras, 2020; Dhakal et al., 2016; Durand, Franks, & Hosken, 2003; Jeske, Zannini, & Arendt, 2017; Mäkinen, Uniacke-Lowe, O’Mahony, & Arendt, 2015; Walther et al., 2022; Zheng, Zhang, Lin, & McClements, 2019). LD technique provides the volume-weighted size (D 4,3 ) of milk particles in solution since milk is a liquid (Grossmann et al., 2021). However, since this technique relies on the Mie theory for size determination in solution, the refractive index of the particles being measured must be known (Grossmann et al., 2021). Additionally, the Mie theory assumes that the molecular species in plant-based milks are non-interactive spheres, which is not the case. Plant-based milk components often aggregate in solution (Grossmann et al., 2021), as demonstrated by the aggregation of fat under heat and pressure in almond milk (Dhakal et al., 2016). Modern LD instruments also provide the span, which indicates the spread of the data by showing how far size percentiles (D 90 ) are from D 10 , normalised at D 50 . These size distributions are usually calculated using the volume-weighted size (D 4,3 ), which is sensitive to the presence of larger particles (Grossmann et al., 2021). This makes D 4,3 particularly suitable for plant milk products, as they often contain large plant material that survives homogenisation and ultra-high temperature treatments during processing. Some researchers have reported the area-based hydrodynamic size (D 3,2 ) for plant milk products (Durand et al., 2003; Mäkinen et al., 2015). While oat and rice milks were sometimes found to have particles with D 3,2 sizes greater than 1 µm, other studies reported particles sizes below 0.5 µm for these products (Mäkinen et al., 2015). Nevertheless, D 4,3 is the most reported value, as it highlights particle sizes in plant milks effectively. LD has been widely used to determine the particle sizes of long-life soy, oat, quinoa, almond, hemp, hazelnut and rice milks (Dhakal et al., 2016; Gallier, Gordon, & Singh, 2012; Ho et al., 2023; Jeske et al., 2017; Mäkinen et al., 2015; Zheng et al., 2019). In some studies, LD was also employed to assess the stability of commercially available plant milks through centrifugation and subsequent particle size measurements (Durand et al., 2003). This research showed that plant milks often exhibit bi- or trimodal particle size distributions. Unlike oat milk, soy and bovine milk contained a smaller proportion of large particles (Durand et al., 2003). The average D 50 particle sizes for bovine, soy, oat and rice milks were found to be approximately 0.43 µm; 0.47 µm; 6.11 µm and 6.64 µm, respectively. Notably, oat and rice milk had larger particles sizes compared to soymilk. DLS has been less commonly used for particle size determination in plant-based milk products. One limitation of DLS instruments is their size range, which typically does not exceed 10 µm. Since plant-based milk products contain large particles, often derived from plant materials, DLS may not be ideal. In contrast, LD tis better suited for measuring larger particles. Moreover, DLS measurements can be influenced by the presence of dust particles, which scatter light in the forward direction and compromise measurement accuracy. Nevertheless, DLS has been used to determine the particle size distribution of soymilk stabilised with konjac glucomannan, a natural polysaccharide (An, Kang, & Tian, 2019). The results showed a bimodal distribution of soymilk particles with particles under 1µm in size. This is a smaller than the particle sizes typically reported using LD for soymilk. Particle sizes in plant-based milk products vary widely due to factors such as the geographical origin of raw materials, plant species, rainfall patterns, and processing parameters, including homogenisation and treatment conditions. This variability contributes to the differences in reported particle sizes across different plant-based milks globally. As a bulk intensity-based technique, DLS is a useful for detecting particle aggregation, which may indicate emulsion instability on plant-based milks. Key parameters reported in DLS studies are the average hydrodynamic size (D z or D H ) and the polydispersity index (PDI), which reflects the particle size and its variation within a population. For instance, DLS was employed to study the effect of sonication on chickpea milk (Vallath & Shanmugam, 2022), highlighting its capability to detect agglomeration or aggregation in solution. In this study, sonication reduced the particle size, but aging led to an increase. For example, when chickpea milk was sonicated for 6 min at 195 W and subsequently aged, the average particle size (D 50 ) increased from 0.97 ± 0.01 µm to 1.26 ± 00 µm by day 14, indicating aggregation. DLS has also been used to measure the hydrodynamic size of plant components in solution, such as proteins glycinin, a soy globulin (Pizones Ruiz-Henestrosa, Martinez, Patino, & Pilosof, 2012). In this study, intensity distributions were converted to volume distributions to mitigate the scattering intensity bias introduced by large particles. The study found that the average hydrodynamic sizes (D H ) of glycinin components (3S, 7S, 11S and 15S) ranged between 6 and 21 nm, which are relatively small compared to the bulk intensity measurements of plant-based milk products. Similarly, to LD, DLS technique assumes all particles are spherical. This assumption can result in the underestimation of hydrodynamic size for plant milk components, as these particles often aggregate to form larger agglomerates. Therefore, to better evaluate such aggregations, microscopy is a complementary and effective technique for analysing plant-based milks. Particle size, shape and aggregation: microscopy When microscopy is used in plant milk research, the research objective determines the most suitable microscopic technique. For measuring particle size, light scattering techniques such as DLS and LD are commonly used. However, for studying morphology, shape, molecular interactions, and surface structure, CLSM and electron microscopy (EM) are the techniques of choice. Since plant milk components are in solution, they tend to aggregate due to unavoidable physical forces. To stabilise these emulsion systems, plant-based milk manufacturers often incorporate natural and artificial emulsifiers, such as proteins and gums. These emulsifiers act as stabiliser, ensuring that milk particles repeal each other, resulting in a stable plant milk product. Microscopy techniques, particularly CLMS with fluorescent labelling, have been used to analyse the microstructure of the major components in plant-based milks (An, 2019; Dhakal et al., 2016; Stephanie Jeske, Juergen Bez, Elke K Arendt, & Emanuele Zannini, 2019; Krongsin, Gamonpilas, Methacanon, Panya, & Goh, 2015a). Researchers have co-stained proteins and fat molecules in plant-based milks, observing uniform protein distribution in soymilk emulsions (An, 2019). Additionally, a “protein halo” surrounding fat globules has been identified in raw almond milk (Dhakal et al., 2016). However, microscopy has inherent limitations in resolution (Gray, 2009). The results obtained depend on factors such as sample preparation, the staining efficiency of fluorescent dyes, the field of view, and the expertise required to produce repeatable results. Despite these limitations, microscopy complements light scattering techniques such as DLS and LD, offering valuable insights into the aggregation state, morphology, and structural properties of plant-based milk components including fats, proteins and carbohydrates (Grossmann et al., 2021). Optical microscopy, when combined with fluorescent dyes, enables the localisation of proteins, fats, carbohydrates, and their assemblies in solution. In contrast, EM provides detailed information about the size and structure of these assemblies. Depending on the research objective, some studies have utilised optical microscopy without fluorescent dyes (Durand et al., 2003). However, the most widely used method for determining microstructures involves fluorescent dyes coupled with CLSM. This technique has been used to characterise lipid molecules (fats), proteins and carbohydrates in plant-based milks and their emulsions (Bernat, Cháfer, Rodríguez-García, Chiralt, & González-Martínez, 2015; Gallier et al., 2012; Stephanie Jeske, Juergen Bez, Elke K. Arendt, & Emanuele Zannini, 2019; Krongsin et al., 2015a; Krongsin, Gamonpilas, Methacanon, Panya, & Goh, 2015b). Although limited by resolution, around half the wavelength of the light used (Smolyaninov, 2008), optical microscopy with fluorescent labelling has proven to be an easy method for investigating molecular interactions among plant milk components. For example, CLSM with Nile Red and Fast Green FCF has been used to stain lipids and proteins, respectively, in almond oleosomes within almond milk (Gallier et al., 2012). This research also showed the co-localisation of glycoproteins and glycolipids on the surface of oil bodies by co-staining with conjugated lectins, such as The Alexa Fluor® 488 conjugates of wheat germ agglutinin (WGA) and concanavalin A (Gallier et al., 2012). These findings revealed that proteins form a halo around the lipid core by attaching its surface, with glycoproteins contributing to the proteins on the almond oil body surface. Unlike bovine milk, which contains casein micelles, almond milk lacks protein bodies, as its free proteins are dispersed in solution, as shown by the CLSM images (Gallier et al., 2012). Similarly, other studies using CLMS have reported a thin protein film surrounding oil bodies in almond milk (Dhakal et al., 2016). CLSM has also been employed to assess the effects of pulsed electric field (PEF) treatment on the particle size of almond milk (Manzoor et al., 2019). Visual examination showed a reduction in particle size after PEF treatment, a finding further confirmed using LD. This study detected protein bridges in almond milk, with CLSM visually identifying their presence. Dual staining experiments also suggested that the oleosomes were circular in shape. Scale bars on the CLSM images allow for particle size estimation, either visually or through software like FIJI (ImageJ), suggesting that almond oil bodies are typically under 10 µm in size (Dhakal et al., 2016; Gallier et al., 2012). CLSM has also been used to visualise agglomeration of proteins in soymilk stabilised with the fibre konjac glucomannan (KGM) (An et al., 2019). A uniform distribution of proteins was observed, with noticeable protein agglomeration at 80 % v/v soymilk with KGM. In this study, proteins were stained with Nile red, while KGM remained unstained, demonstrating the ability CLSM to perform both positive and negative staining for specific molecular components in plant-based milks. EM techniques have a higher resolution than optical microscopy. These techniques operates by using a beam of electrons instead of light, either passing through the sample in transmission electron microscopy (TEM) or bouncing off the surface in scanning electron microscopy (SEM), to reveal detailed microstructural information (Grossmann et al., 2021). For example, Manzoor et al. (Manzoor et al., 2019) used EM to study almond milks, revealing that they have nonporous, smooth-surfaced, and sharp-edged particles. This highlights SEM’s capability to elucidate particle structure and morphology in plant milk products. In addition, a combination of light microscopy and TEM has been utilised to image various plant-based milk products, including soy, oat, quinoa and rice (Mäkinen et al., 2015). This study showed that soy and quinoa milks contained uniformly sized oil-based particles along with larger fragments, potentially plant material, while rice milk lacked these large fragments. However, it is important to note that SEM sample preparation involves numerous dehydration steps, which can disrupt the molecular structures of milk components (Grossmann et al., 2021). Therefore, careful consideration is necessary when selecting characterisation techniques for plant milk products, as the choice should align with the specific goals of the research. Surface charge: ζ-potential Plant-based milk manufacturers aim to produce products with prolonged shelf life and optimal performance for applications such cooking, baking addition to hot beverages like tea or coffee. Stability in these applications is influenced by the surface charge of milks particles, quantified as zeta potential (ζ-potential). ζ-potential represents the charge at the electrical double layer of nanoparticles in solution (Clogston & Patri, 2011; Delgado, González-Caballero, Hunter, Koopal, & Lyklema, 2007; Lyklema, 2005) and serves as an indicator of colloidal stability. Stability depends heavily on the ionic environment (pH and ionic strength), and significant changes to this environment can disrupt the emulsion stability, affecting consumer satisfaction. Electrophoretic light scattering (ELS) is commonly used to measure the ζ-potential of plant-based milk products. A ζ-potential greater than ±30 mV indicates particle stability, while values between –30 mV and +30 mV suggest susceptible to emulsion instability (Honary & Zahir, 2013). Since ζ-potential depends on pH and dilution (necessary for measurement), reporting pH of the dispersant is crucial for meaningful interpretation. In most cases, research has shown that plant-based milk particles are typically negatively charged at neutral or near-neutral pH. For instance, soymilk particles had a ζ-potential of less than −30 mV (An, 2019), almond oil bodies in almond milk exhibited a ζ-potential of −29.9 ± 1.99, and soy protein isolate showed a ζ-potential of approximately −20 mV (Krongsin et al., 2015a). Similarly, hazelnut milk particles displayed ζ-potential values ranging between −29 to −33 mV (Şen & Okur, 2022), while unsonicated chickpea milk emulsions were reported to have a ζ-potential of −25.8 mV (An, 2019; Gallier et al., 2012; Krongsin et al., 2015a; Şen & Okur, 2022; Vallath & Shanmugam, 2022) Studies have also used ζ-potential to evaluate the impact of stabilisers on plant milk stability. For example, KGM was found to improve soymilk stability, showing more negative ζ-potential at higher concentrations (An et al., 2019). Similarly, pH-driven curcumin loading in soymilk revealed emulsion instability between pH 2 and 5, with a shift in ζ-potential from −39.8 ± 2.6 mV at neutral pH to positive values near the isoelectric point (pI) at pH 3.6 (Zheng et al., 2019). Such analyses are essential for understanding the pH stability range and the pI, where proteins aggregate and precipitate, helping manufacturers optimise formulations for acidic environments like coffee. Additionally, studies on stabilisers like pomelo pectin demonstrated improved stability of soy protein isolates at acidic pH (Krongsin et al., 2015a) Overall, understanding the surface charge of plant-milk particles aids the industry in improving formulations, enhancing stability, and meeting consumer demands for reliable, high-quality products. Separation plant-based milk components As discussed above a range of analyses on the whole milk product can be undertaken. However, when more detailed understanding of the molecular composition of the milks is required, different approaches are required. Some chemical methods can be applied to the whole milk but only measure chosen components, other methods require separation by mechanical methods and/or chromatographic methods prior to chemical or further chromatographic analysis. DD MMMM YYYY \acceptedDD MMMM YYYY Centrifugation and filtration To efficiently analyse plant milk components, mechanical separation techniques such as centrifugation and filtration are often used prior to chemical analysis. Centrifugation is a mechanical separation method that uses centripetal force to induce sedimentation and separate milk components of different densities, while filtration separates larger particles from smaller ones. Depending on the aim of the experiment, some researchers have used only centrifugation as a technique to separate molecular components of long-life soy, rice and oat milk products. Durand and co-workers (Durand et al., 2003) observed a size separation between the floating and sinking fractions of commercially available plant milks. In soymilk, the floating portion contained larger particles than the sinking portion, while in oat milk, the sinking portion had larger particles (Durand et al., 2003; Mäkinen et al., 2015). (Jeske et al., 2017) used 0.25 µm nylon filters for filtration, followed by dilution of the filtrate prior to chromatographic analysis for mono- and disaccharide content (sugar analysis) in commercially available plant milks. The material of the filters used may affect the recovery of desired plant-milk components, as nitrocellulose and polyvinylidene (PVDF) are known to bind to proteins (Tovey & Baldo, 1989), potentially reducing their recovery. Other membrane materials, such as cellulose, have been used for tryptophan analysis in commercially available plant-based milks (Walther et al., 2022). The membrane pore size of filters used for sample preparation is also an important consideration. For instance, some researchers have used 0.45 µm filters (Mäkinen et al., 2015) while others have used 0.25 µm for sugar analysis (Jeske et al., 2017). Since mechanical separation is an incomplete technique, it is often complemented with chemical methods to reduce matrix effects. Chemical extraction Chemical extraction techniques are often utilised for plant-based milk separation to isolate types of molecules (fats versus sugars versus proteins) with minimal interference from the sample matrix (Mäkinen et al., 2015; Walther et al., 2022) to ensure distinct responses when milk molecular species are analysed. Chemical separation for plant-based milks has employed the use of organic/inorganic solvents, buffers and acids to extract their molecular components. The aim of chemical separation is to minimise matrix interference, but it assumes that the matrix allows for 100% extraction. Researchers have used organic solvents such 80 % v/v ethanol:water for the extraction of oligosaccharides in long-life soy and almond milks (Huang, Paviani, Fukagawa, Phillips, & Barile, 2023). Others have used methanol, and acetonitrile (Thanavanich, Phuangsaijai, Thiraphatchotiphum, Theanjumpol, & Kittiwachana, 2022) to extract sugars or simply diluted the milk with water prior to sugar analysis by HPLC-RID (high performance liquid chromatography refractive index detection, see below) (Jeske et al., 2017). Additionally, Carrez clarification using Carrez I (potassium hexacyanoferrate) and Carrez II (zinc acetate dihydrate) ( Fig. 4 ) has been used to remove proteins and fats for sugar analysis of long-life plant milk products via HPAEC-PAD (high performance anion exchange chromatography with pulse ampometric detection, see below) and HPLC-RID (Huang et al., 2023; Mäkinen et al., 2015). Following Carrez clarification, filtration and dilution, it was concluded that the milk samples could be directly injected into the HPAEC. However, if ethanol (or other organic solvents) had been used, they must be removed, and the samples dissolved in water prior to analysis (Huang et al., 2023). Liquid-liquid extraction of plant-based milk components is a chemical separation technique that can isolate milk components based on their solubility in organic solvents. This extraction can be performed using methanol/chloroform extraction. This technique uses two immiscible organic solvents (methanol and chloroform) to extract the molecular species of interest into the solvent which the molecules are most soluble in. This extraction method has not been well explored for plant-based milks; however, it has been used to extract proteins for understanding the protein-bound gycome analysis of wallaby milk (Wongtrakul-Kish et al., 2013). This separation technique forms a proteinaceous disc in the interface of the aqueous (methanol/water) top layer and organic (chloroform) bottom layer ( Fig. 5 ). Since carbohydrates are water soluble molecules, they would possibly migrate to the top of the tube (aqueous layer) while fats will be found in the organic (bottom) layer separated by the proteinaceous disc. However, because of the varying solubilities of milk molecular components, some hydrophilic proteins may also be present in the aqueous fraction, which is located at the top of the tube. Additionally, due to molecular interactions between milk molecules, glycoproteins and glycolipids may migrate to either solvent phase, potentially affecting their recovery. Therefore, careful consideration must be placed when selecting the appropriate chemical separation technique for optimum plant-milk component content. DD MMMM YYYY \acceptedDD MMMM YYYY Chromatographic separation The column Analytical and/or preparative chromatography separation techniques, sometimes following mechanical or chemical extraction or chemical derivatisation, are often utilised for plant-based milk to separate and sometimes quantify individual components(Mäkinen et al., 2015; Walther et al., 2022). Separated components may they be analysed chemically. The separation of plant-based milk components can be either complete or partial, as shown in Fig. 6 . Complete separation occurs when components 1, 2 and 3 are spatially isolated from each other ( Fig. 6 A ), whereas partial separation occurs when component 1 and 2 remain mixed, while component 3 is spatially isolated from the mixture ( Fig. 6 B ) (Skoog & Skoog, 2014). Chromatography is the most used method for separation and analysis of plant milk components. Key components of chromatographic systems are sample introduction, column containing a stationary phase through which the mobile phase passes and a detector. It is a technique that is based on one phase being held in place (stationary phase) while the other phase is moving (mobile phase) (Coskun, 2016). A schematic of a high-pressure liquid chromatography (HPLC) is shown in Fig. 7 . Alternatively, if the mobile phase used is an inert gas such as helium, it is referred to as gas chromatography (GC). Both systems may be used in preparative or analytical chromatography. The analytes of interest adsorb on the stationary phase and are sequentially eluted based on their ionic interaction with the column stationary phase. GC analysis requires volatile analytes. Plant-based milk sugar analysis may commence with O-methyl hydroxylamine hydroxide in pyridine for methoxymation, followed by trimethylsylation using derivatisation reagents like N-trimethyl-N-methyl trifluoroacetamide containing 1% of trimethylchlorosilane (MSTFA + 1% TMCS) (Jariyasopit et al., 2021). This process produces more volatile, derivatised sugars that can be separated on a GC column. Usually, hydrochloric acid (HCl), acetyl chloride (CH 3 COCl), sulfuric acid (H 2 SO 4 ), or boron trifluoride (BF 3 ) (Chiu & Kuo, 2020) with heating are used as derivatisation reagents for fatty acids prior to GC analysis. Analytical HPLC columns are either hydrophobic or hydrophilic, with lengths up to 250 mm and internal diameters ranging from 3.0 to 4.6 mm. Analytical columns are typically packed with porous silica particles, typically 1.8, 3.5, or 5 µm in diameter, while semi-preparative column uses 7 µm particles. Standard analytical silica-based HPLC columns can withstand pump pressures up to 400 bar. When the column stationary phase is hydrophilic (polar) and the eluent is hydrophobic (non-polar) such as hexane, the separation is referred to as normal-phase HPLC. The separation efficiency of normal phase HPLC is generally poor. So, both analytical and preparative silica may be derivatised to alter their interaction with analytes. The result is so-called reverse-phase HPLC, where the stationary phase is hydrophobic, and the mobile phase is polar (e.g., methanol/acetonitrile). Reverse-phase HPLC is commonly used for the analysis of major milk components (protein, lipid, and sugar). For sugar separation, the mobile phase typically consists of acetonitrile, methanol, and water (Jeske et al., 2017; Mäkinen et al., 2015; Sharma, Rajput, Poonam, Dogra, & Tomar, 2009). Protein content analysis using reverse phase HPLC has been used to determine the amino acid sequence profile of plant-based milks. In this work, amino acid sample preparation often involves hydrolysing the proteins, followed by the amino acids derivatisation using Acc-Q-Tag Ultra reagent prior to analytical reverse phase HPLC analysis (Walther et al., 2022). Neither normal nor reverse phase columns may give sufficient separation for some milk analytes. High performance anion exchange chromatography (HPAEC) has proved a versatile separation technique, which is used for a range of anionic molecules, including organic acids, amino acids, nucleotides, peptides, and sugars. The CarboPac PA1 column is made of 10 µm beads, surface-sulfonated and coated with smaller beads electrostatically bound to the larger ones, containing a quaternary amine group that gives the surface a positive charge. Other strong anion-exchange columns such as CarboPac PA100 and CarboPac PA200, are also be used for oligosaccharides separation at higher sodium hydroxide or sodium acetate concentrations (Corradini, Cavazza, & Bignardi, 2012). For sugar analysis, polymeric anion-exchange columns such as Dionex CarboPac PA1 and PA10 are typically used for mono- and disaccharide analysis (Corradini et al., 2012). Such columns offer high-resolution separation based on hydroxyl group number. Unlike HPLC, which can withstand higher pressures, HPAEC requires less pressure for column separation. Size exclusion chromatography (SEC), while less commonly used in plant milk analysis due to its lower resolution, has been employed to determine oat β-glucan size in two different cultivars (Skendi, Biliaderis, Lazaridou, & Izydorczyk, 2003). Recently, HPAEC-PAD has been used to identify and quantify oligosaccharides like raffinose, stachyose and verbascose in commercially available soy/almond milk products and soy flour, achieving low limits of detection (LOD) of 1.2 µg/L for raffinose, 1.3 µg/L for stachyose, and 3.4 µg/L for verbascose (Huang et al., 2023). The detector A key feature of any chromatography system is the detector. GC detectors for plant-based milk analysis are generally flame ionisation (FID) which is sensitive but provides no identification other than the retention time (analyte time on column) or mass spectrometric (MS) which provides an MS finger print for different analytes with about 0.1 Da resolution (Skoog & Skoog, 2014). Absorbance spectroscopy is most commonly used detector for HPLC but cannot be used for sugars which do not have a chromophore in the far or near UV or the visible region of the spectrum. So, after column separation, plant-based milk components are usually detected using refractive index detectors (RID) or pulsed amperometric detection (PAD) for sugars, ultraviolet (UV) detection for proteins/amino acids or mass spectrometry (MS). Since sugars in plant-based milks do not strongly absorb in the UV range (Liu, Chen, & Lin, 2002), RID is commonly used (Jeske et al., 2017; Mäkinen et al., 2015). However, RIDs have low sensitivity, cannot be used with gradient mobile phases, and are affected by pump pressure fluctuations, temperature changes, and variations in mobile phase composition or flow rate (Cataldi, Angelotti, & Bianco, 2003; Dolan, 2012) (Hirata, Kawaguchi, & Funada, 1996), potentially leading to an underestimation of sugar content. HPAEC-PAD, in contrast, is more sensitive, offering high-resolution separation based on hydroxyl group number and an electrochemical detector that can detect sugars at picomole concentrations. Content analysis of plant-based milks The most common plant-based milk analysis methods are outlined below. Total protein content Total nitrogen is used to estimate protein content using the Kjeldahl method (Jeske et al., 2017; Mäkinen et al., 2015; Walther et al., 2022). This method involves digestion of the milk using a strong acid such as sulfuric acid so that the nitrogen released is quantified by titration (Hayes, 2020). Following determination of the total nitrogen, a specific conversion factor is needed to convert the measured nitrogen into total protein content. Although each amino acid has one amide group, some amino acids have 2 nitrogens, and the residue molecular masses vary from 75 to 204 mass units, so the average amino acid mass varies depending on the proteins in the milk. The conversion factor approved by the World Health Organisation (WHO) and the Food and Agriculture Organization (FAO) of the United Nations for plant-based milks is approximately 5.6. Some researchers have used a conversion factor of 5.6 (Walther et al., 2022), while others have used 5.75 for rice milk and 5.95 for soy, oat and quinoa milks (Mäkinen et al., 2015). The choice of conversion factor can introduce variations of over 6% in the reported protein content. Some researchers have estimated protein content in plant-based milks using the Bradford assay (Dhakal et al., 2016) which involves covalently binding the dye Coomassie Brilliant Blue G-250 to basic and aromatic amino acid residues of proteins and measuring the colour change. Researchers have reported that soymilk contains the highest protein content among plant-based milks. The protein content of commercially available soymilk has been reported as 2.95 ± 0.07 g/100g (Mäkinen et al., 2015), 3.70 ± 0.03 g/100g for plain UHT Soya drink, 2.61 ± 0.13 g/100g for Soya Original drink, 3.16 ± 0.32 g/100g for Soya Organic, Wholebean and 2.61 ± 0.13g/100g for Soya original (Jeske et al., 2017). Other studies reported 3.78 g/100g for commercially available soy drink (Walther et al., 2022). In this same study, soymilk was found to be superior in both essential and non-essential amino acid content compared to almond, cashew, coconut, oat, rice and other commercially available plant milks (Walther et al., 2022). In contrast, rice milk has been shown to have lower protein content, with values of 1.7 g/100g (Walther et al., 2022), 0.32 ± 0.04 g/100g for Organic rice drink natural, and 0.07 ± 0.00 g/100g for Organic brown rice drink (Jeske et al., 2017). Another commercially available rice milk contained 0.07 ± 0.02 g/100g (Mäkinen et al., 2015). Fat content Crude fat content has been determined using the Weibull-Stoldt method gravimetrically through fat and protein hydrolysis and Soxhlet extraction (Jeske et al., 2017; Standard, 2005) or the Gerber method (Mäkinen et al., 2015). The latter involves separating milk fat from proteins using sulfuric acid followed by direct fat measurement using a calibrated butyrometer. Alternatively, high resolution lipid analysis may be performed using gas chromatography (GC) after fatty acid derivatisation using a boron trifluoride/methanol mixture (usually 14 % w/v) with flame ionisation detection (FID) (Walther et al., 2022) or mass spectrometry (MS). Fat content also varies between different plant milk types and manufacturers, as shown in Table 1 . For example, The good little Cook almond milk contains 4.40 ± 0.11 g/100g, while Alpro Original almond milk has 1.18 ± 0.05 g/100g, Provamel Organic almond milk contains 3.69 ± 0.11 g/100g, and The Good Little Cook Carob almond milk has 3.35 ± 1.73 g/100 (Jeske et al., 2017) . In the same study, Vitariz Organic Rice Drink, Oatly Oat Drink, and Alpro Coconut Original were reported to have lower fat content, with values of at 0.85 ± 0.06 g/100g, 0.38 ± 0.06 g/100g, and 0.84 ± 0.00 g/100g, respectively (Jeske et al., 2017). Other studies have reported quinoa milk to contain the highest fat content, at 2.40 ± 0.10 g/100g, with a hierarchy of quinoa>soy>rice>oat (Mäkinen et al., 2015). Additionally, commercially available Hemp milk has been reported to contain a higher fat content of 3.26 g/100g, while Coconut milk has a lower fat content of 1.02 g/100g (Walther et al., 2022). Mineral and total ash content Mineral content of milks is typically determined using Inductively Coupled Plasma (ICP) Optical Emission spectroscopy (OES) or mass spectrometry following nitric acid microwave digestion (Walther et al., 2022). Minerals in plant-based milks can either be intrinsic to the raw plant material or added as fortifying agents to improve the nutritional profile. However, the mineral content of plant milks have been found to be generally inferior compared to cow’s milk (Walther et al., 2022). If total ash is desired for research purposes, it is normally determined by burning plant milks in a muffle furnace in a crucible prior to measurement of mass of residue after cooling (Dhakal et al., 2016; Mäkinen et al., 2015). Carbohydrate content Carbohydrate analysis in plant milk products has been performed using ultraviolet-visible (UV-Vis) spectrophotometry assays with enzymes (Walther et al., 2022). This process involves enzymatically hydrolysing long-chain polysaccharides into simpler sugars, which are then complexed with sugar-specific visible light-absorbing compounds. These sugars are quantified colorimetrically using a sugar standard calibration curve based (usually) on glucose. Chemical colorimetric assays, such as the anthrone or phenol-sulfuric acid test (Katoch & Katoch, 2011), may also be utilised. The anthrone test works by dehydrating pentoses and hexoses into furfural and hydroxyfurfural, respectively, which then complex with the anthrone reagent (9,10-dihydro-9-oxoanthracene), resulting in a blueish-green complex whose absorbance can be measured spectrophotometrically (Katoch & Katoch, 2011). This sugar estimation technique, based on glucose, may led to an underestimation of the sugar content in plant-based milks, as they contain more than just glucose. Although highly unusual, the sugar content of plant-based milks can be determined using more accurate and robust analytical techniques, such as the chromatography methods outlined above including high-performance anion exchange chromatography with pulsed amperometry detector (HPAEC-PAD) or size exclusion or reverse phase chromatography with refractive index detectors (HPLC-RID). In most uses of these techniques, the total sugar content is estimated by integrating the areas under the curves for the individual sugars (e.g. glucose, sucrose, maltose, fructose) analysed through chromatography. However, they are also useful for quantifying the different sugars in a product if a calibration standard is available. In terms of sugar (i.e. small carbohydrate molecules) content, the most abundant sugars in plant milk products are sucrose and glucose (Walther et al., 2022), with maltose also present in some plant milk products (Jeske et al., 2017). Additional glucose may be formed enzymatically during production through starch hydrolysis (Sethi et al., 2016), while sucrose is typically added during manufacturing as a sweetener, thereby enhancing the taste of the milk. In terms of total carbohydrate content, long-life oat and rice milk have been shown to have the highest concentration, with 3.67 and 4.82 g/100g, respectively (Walther et al., 2022). This high carbohydrate content contributes to the mouthfeel and creamy texture of these milks when consumed. Conclusion and Future Perspectives Plant-based milk products provide an alternative source of proteins, fats, carbohydrates, and minerals for many consumers. Before these products are releases to the market, rigorous analytical tests are required to meet consumer expectations and regulatory standards. Evidence indicates molecular interactions between proteins and fats, with proteins forming a “halo” around oil bodies and sugars attaching to their surfaces. These interactions have been observed using optical microscopy, specifically CLSM coupled with fluorescent dyes. However, due to resolution limitations and research goals, microscopy has paired with particle sizing techniques, particularly LD, which remains the most utilised particle sizing technique in plant-based milks. It remains unclear whether these molecular interactions in plant-based milks affect their sensory and functional properties. Animal milks such as bovine contain calcium in intermolecular structures known as casein micelles. Therefore, calcium is integrated into these structures which makes it bioavailable when consumed. On the other hand, plant-based milks have low intrinsic calcium and therefore manufacturers add this mineral to enhance their nutritional. Additionally, manufacturers typically add calcium at concentrations low enough to avoid intermolecular interactions with proteins which would precipitate them affecting product quality. Nevertheless, it is unclear which intermolecular structures are formed within plant-based milks when calcium is added to them, and whether these products produce molecular structures such as casein micelles, which indirectly affects calcium bioavailability. Currently, particle size and surface charge measurements (e.g., ζ-potential via electrophoretic light scattering, ELS) are predominantly used in research settings rather than for QC in production. Variations in particle size and ζ-potential have been documented, influenced by environmental factors such as raw material origin, rainfall patterns, and geographical location. Manufacturers analyse plant-based milk components to determine their content on NIP using methods that often involve the disruption of intermolecular structures (e.g., acid hydrolysis or enzymatic treatments). These approaches, however, may underestimate component content. A summary of observed molecular interactions and the analytical techniques used in plant-based milks is shown in Fig. 8 . Importantly, there is limited evidence regarding interactions between proteins, fats, carbohydrates, and minerals, such as calcium, a critical nutrient for many milk consumers. Molecular component content in plant-based milks has been shown to vary significantly across products, even when derived from the same plant source, due to the addition of various components during production. While total sugar content is often reported on the NIP, it is evident that these values depend on the specific sugars analysed. For example, the lower sensitivity of RID and UV-Visible detection compared to PAD can lead to underestimation of sugar content. The use of highly sensitive detectors, such as PAD or MS, warrants consideration for such analyses. Efforts must therefore focus on the development of robust, accurate, and reproducible analytical techniques to characterise plant-based milk components comprehensively. These advancements will support manufacturers in delivering high-quality products that meet consumer expectations and regulatory requirements. Fig. 1. Environmental impact of dairy and plant-based milks. The cost to the environment for producing 1 litre of milk, from farm to retail, as measured by the four environmental impact indicators: Land use, Greenhouse gas emissions, Freshwater use, and Eutrophication. Taken from Our world in Data (Poore & Nemecek, 2018; Ritchie, 2022). Fig. 2. An illustration of β-amylase action on amylopectin (part of starch molecule). Idea extracted from α-amylase action described by (Visvanathan et al., 2020). Fig. 3. Structure of (A) bovine milk fat globule and (B) oleosomes in plant-based milks (Created on https://BioRender.com) (Nikiforidis, 2019; Singh & Gallier, 2017). Fig. 4. A simplified illustration of the formation of zinc hexacyanoferrate salt from potassium hexacyanoferrate (Carrez I) and zinc acetate (Carrez II) solutions. This complex then binds to proteins through adsorption and precipitates them out of solution. Precipitated insoluble material is removed by centrifugation or filtration (Carrez, 1908). Fig. 5. Photograph of (A) Almond milk 1 and (B) almond milk 2 after methanol and chloroform extraction. A protein disc forms at the interface of the chloroform and aqueous methanol layer. Fig. 6. Separation phenomena of ( A ), a mixture of three molecular species is completely separated to form isolated fractions 1, 2 and 3 such as in chromatography of plant milk molecular components while in ( B ) molecular species 1 and 2 are still molecularly intact while species 3 is completely isolated such as in centrifugation or filtration (mechanical separation) of plant-milk components. Adapted from (Skoog & Skoog, 2014). Fig. 7. A schematic of the high pressure liquid chromatography (HPLC) system (Czaplicki, 2013). Created on https://BioRender.com Fig. 8. A summary of analytical techniques and evidence of molecular interactions used in plant-based milk analysis from a total and component milk analysis. Milk carton created in https://BioRender.com Acknowledgement This work was supported by the Australian Research Council Industrial Transformation Training Centre for Facilitated Advancement of Australia’s Bioactives (Grant IC210100040) and the Research Attraction and Acceleration Program funding from the Office of the Chief Scientist and Engineer, Investment NSW. Supplementary Material File (image2.emf) Download 213.32 KB References 1. https://ourworldindata.org/environmental-impact-milks An, K. (2019). Stabilization of soy milk using konjac glucomannan. Emirates journal of food and agriculture , 526-534. An, K., Kang, H., & Tian, D. (2019). Stabilization of soy milk using konjac glucomannan. Emirates journal of food and agriculture, 31 (7), 526-534. S. M., Connor, S. L., Sexton, G., & Connor, W. E. (1993). Differences in coronary mortality can be explained by differences in cholesterol and saturated fat intakes in 40 countries but not in France and Finland. A paradox. Circulation, 88 (6), 2771-2779. W., & Knoop, A. (1959). ber die Morphologie der Milchsekretion Licht- und elektronenmikroskopische Studien an der Milchdr se der Ratte. Cell and tissue research, 49 (3), 344-388. N., Cháfer, M., Rodríguez-García, J., Chiralt, A., & González-Martínez, C. (2015). Effect of high pressure homogenisation and heat treatment on physical properties and stability of almond and hazelnut milks. Food science & technology, 62 (1), 488-496. S., & Gudmand-Høyer, E. (1988). Incidence and Clinical Significance of Lactose Malabsorption in Adult Coeliac Disease. Scandinavian journal of gastroenterology, 23 (4), 484-488. M. (1908). Le ferrocyanure de potassium et l’acétate de zinc comme agents de défécation des urines. Annales de chimie analytique, 13 , 97-101. Cataldi, T. R. I., Angelotti, M., & Bianco, G. (2003). Determination of mono- and disaccharides in milk and milk products by high-performance anion-exchange chromatography with pulsed amperometric detection. Analytica chimica acta, 485 (1), 43-49. H.-H., & Kuo, C.-H. (2020). Gas chromatography-mass spectrometry-based analytical strategies for fatty acid analysis in biological samples. Yàowu shi͡p︡in fenxi, 28 (1), 60-73. J. D., & Patri, A. K. (2011). Zeta Potential Measurement. Characterization of Nanoparticles Intended for Drug Delivery, 697 , 63-70. C., Cavazza, A., & Bignardi, C. (2012). High-Performance Anion-Exchange Chromatography Coupled with Pulsed Electrochemical Detection as a Powerful Tool to Evaluate Carbohydrates of Food Interest: Principles and Applications. International Journal of Carbohydrate Chemistry, 2012 , 1-13. O. (2016). Separation techniques: Chromatography. Northern clinics of Istanbul, 3 (2), 156-160. S. (2013). Chromatography in Bioactivity Analysis of Compounds. In: IntechOpen.Delgado, A. V., González-Caballero, F., Hunter, R. J., Koopal, L. K., & Lyklema, J. (2007). Measurement and interpretation of electrokinetic phenomena. Journal of colloid and interface science, 309 (2), 194-224. A., Deora, N. S., & Mishra, H. N. (2014). Optimization of enzymatic production process of oat milk using response surface methodology. Food and bioprocess technology, 7 , 610-618. Devnani, B., Ong, L., Kentish, S., & Gras, S. (2020). Heat induced denaturation, aggregation and gelation of almond proteins in skim and full fat almond milk. Food chemistry, 325 , 126901. Dhakal, S., Giusti, M. M., & Balasubramaniam, V. (2016). Effect of high pressure processing on dispersive and aggregative properties of almond milk. Journal of the science of food and agriculture, 96 (11), 3821-3830. Dolan, J. W. (2012). Avoiding refractive index detector problems. LC-GC North America, 30 (12), 1032-1036. Durand, A., Franks, G. V., & Hosken, R. W. (2003). Particle sizes and stability of UHT bovine, cereal and grain milks. Food Hydrocolloids, 17 (5), 671-678. P. J. (2017). Food Processing Technology - Principles and Practice (4th Edition) (4 ed.). Cambridge: Elsevier.Franke, W. W., Heid, H. W., Grund, C., Winter, S., Freudenstein, C., Schmid, E., . . . Keenan, T. W. (1981). Antibodies to the major insoluble milk fat globule membrane-associated protein: specific location in apical regions of lactating epithelial cells. The Journal of cell biology, 89 (3), 485-494. Freudenstein, C., Keenan, T., Eigel, W., Sasaki, M., Stadler, J., & Franke, W. (1979). Preparation and characterization of the inner coat material associated with fat globule membranes from bovine and human milk. Experimental Cell Research, 118 (2), 277-294. Gallier, S., Gordon, K. C., & Singh, H. (2012). Chemical and structural characterisation of almond oil bodies and bovine milk fat globules. Food chemistry, 132 (4), 1996-2006. Gray, N. (2009). Knowing the limit: Diffraction limit theory. Nature cell biology, 11 (Suppl 1), S8-S8. L., Kinchla, A. J., Nolden, A., & McClements, D. J. (2021). Standardized methods for testing the quality attributes of plant‐based foods: Milk and cream alternatives. Comprehensive Reviews in Food Science and Food Safety, 20 (2), 2206-2233. M. (2020). Measuring protein content in food: An overview of methods. Foods, 9 (10), 1340. Y., Kawaguchi, Y., & Funada, Y. (1996). Refractive Index Detection Using an Ultraviolet Detector with a Capillary Flow Cell in Preparative SFC. Journal of chromatographic science, 34 (1), 58-62. T. M., Bhandari, B. R., & Bansal, N. (2023). Effect of shearing‐induced lipolysis on foaming properties of milk. Journal of the science of food and agriculture, 103 (11), 5312-5321. S., & Zahir, F. (2013). Effect of Zeta Potential on the Properties of Nano-Drug Delivery Systems - A Review (Part 1). Tropical journal of pharmaceutical research, 12 (2). Y.-P., Paviani, B., Fukagawa, N. K., Phillips, K. M., & Barile, D. (2023). Comprehensive oligosaccharide profiling of commercial almond milk, soy milk, and soy flour. Food chemistry, 409 , 135267-135267. S. R., Dupas de Matos, A., Frempomaa Oduro, A., & Hort, J. (2024). Sensory characteristics of plant-based milk alternatives: Product characterisation by consumers and drivers of liking. Food research international, 180 , 114093-114093. C. L., Krajicek, E. J., Savaiano, D. A., & Shin, A. S. (2020). Lactose Intolerance: A Concise Review to Skim the Surface. Mayo Clinic proceedings, 95 (7), 1499-1505. N., Khamsaeng, S., Panya, A., Vinaisuratern, P., Metem, P., Asawalertpanich, W., . . . Khoomrung, S. (2021). Quantitative analysis of nutrient metabolite compositions of retail cow’s milk and milk alternatives in Thailand using GC-MS. Journal of food composition and analysis, 97 , 103785. S., Bez, J., Arendt, E. K., & Zannini, E. (2019). Formation, stability, and sensory characteristics of a lentil-based milk substitute as affected by homogenisation and pasteurisation. European food research & technology, 245 (7), 1519-1531. S., Bez, J., Arendt, E. K., & Zannini, E. (2019). Formation, stability, and sensory characteristics of a lentil-based milk substitute as affected by homogenisation and pasteurisation. European Food Research and Technology, 245 , 1519-1531. Jeske, S., Zannini, E., & Arendt, E. K. (2017). Evaluation of Physicochemical and Glycaemic Properties of Commercial Plant-Based Milk Substitutes. Plant foods for human nutrition (Dordrecht), 72 (1), 26-33. R., & Katoch, R. (2011). Carbohydrate estimations. Analytical Techniques in Biochemistry and Molecular Biology , 67-76. Krongsin, J., Gamonpilas, C., Methacanon, P., Panya, A., & Goh, S. M. (2015a). On the stabilisation of calcium-fortified acidified soy milks by pomelo pectin. Food Hydrocolloids, 50 , 128-136. Krongsin, J., Gamonpilas, C., Methacanon, P., Panya, A., & Goh, S. M. (2015b). On the stabilisation of calcium-fortified acidified soy milks by pomelo pectin. Food Hydrocolloids, 50 , 128-136. B., Domagała, J., & Kasprzak, M. M. (2024). Critical Review of Techniques for Food Emulsion Characterization. Applied sciences, 14 (3), 1069. J. R., Chen, M. J., & Lin, C. W. (2002). Characterization of Polysaccharide and Volatile Compounds Produced by Kefir Grains Grown in Soymilk. Journal of food science, 67 (1), 104-108. J. (2005). Fundamentals of interface and colloid science: soft colloids (Vol. 5): Elsevier.MacGibbon, A. K. H., & Taylor, M. W. (2009). Composition and Structure of Bovine Milk Lipids. In (Vol. 2, pp. 1-42). Boston, MA: Springer US.Mäkinen, O. E., Uniacke-Lowe, T., O’Mahony, J. A., & Arendt, E. K. (2015). Physicochemical and acid gelation properties of commercial UHT-treated plant-based milk substitutes and lactose free bovine milk. Food chemistry, 168 , 630-638. M. F., Ahmad, N., Aadil, R. M., Rahaman, A., Ahmed, Z., Rehman, A., . . . Manzoor, A. (2019). Impact of pulsed electric field on rheological, structural, and physicochemical properties of almond milk. Journal of food process engineering, 42 (8), n/a. I. H., & Keenan, T. W. (1975). Studies on the structure of milk fat globule membrane. The Journal of membrane biology, 21 (1), 65-85. I. H., & Keenan, T. W. (1998). Origin and secretion of milk lipids: The cell biology of milk secretion. Journal of mammary gland biology and neoplasia, 3 (3), 259-273. Mather, I. H., Weber, K., & Keenan, T. W. (1977). Membranes of Mammary Gland. XII. Loosely Associated Proteins and Compositional Heterogeneity of Bovine Milk Fat Globule Membrane. Journal of Dairy Science, 60 (3), 394-402. O. J., & Singh, H. (2009). Physico-chemical properties of milk. In Advanced dairy chemistry (pp. 691-758): Springer.McClements, D. J. (2019). Future foods: How modern science is transforming the way we eat : Springer.McClements, D. J., & Grossmann, L. (2021). The science of plant‐based foods: Constructing next‐generation meat, fish, milk, and egg analogs. Comprehensive Reviews in Food Science and Food Safety, 20 (4), 4049-4100. C. V. (2019). Structure and functions of oleosomes (oil bodies). Advances in colloid and interface science, 274 , 102039. Pizones Ruiz-Henestrosa, V. M., Martinez, M. J., Patino, J. M. R., & Pilosof, A. M. R. (2012). Dynamic Light Scattering Study on the Complex Assembly of Glycinin Soy Globulin in Aqueous Solutions. Journal of the American Oil Chemists’ Society, 89 (7), 1183-1191. J., & Nemecek, T. (2018). Reducing food’s environmental impacts through producers and consumers. Science (American Association for the Advancement of Science), 360 (6392), 987-992. H. (2022). Dairy vs. plant-based milk: what are the environmental impacts? Retrieved from Rotz, C. A., Montes, F., & Chianese, D. S. (2010). The carbon footprint of dairy production systems through partial life cycle assessment. Journal of Dairy Science, 93 (3), 1266-1282. L., & Okur, S. (2022). Effect of hazelnut type, hydrocolloid concentrations and ultrasound applications on physicochemical and sensory characteristics of hazelnut-based milks. Food chemistry, 402 , 134288-134288. S., Tyagi, S. K., & Anurag, R. K. (2016). Plant-based milk alternatives an emerging segment of functional beverages: a review. Journal of food science and technology, 53 (9), 3408-3423. R., Rajput, Y. S., Poonam, Dogra, G., & Tomar, S. K. (2009). Estimation of sugars in milk by HPLC and its application in detection of adulteration of milk with soymilk. International journal of dairy technology, 62 (4), 514-519. A. J., Hamad, I., Abdo, R. W., Qadumii, L., Jaber, A. Y., Surchi, H. S., & Alkelany, S. Z. (2019). Synthesis, characterization, and applications of metal nanoparticles. In Biomaterials and bionanotechnology (pp. 527-612): Elsevier.Singh, H., & Gallier, S. (2017). Nature’s complex emulsion: The fat globules of milk. Food Hydrocolloids, 68 , 81-89. Skendi, A., Biliaderis, C. G., Lazaridou, A., & Izydorczyk, M. S. (2003). Structure and rheological properties of water soluble β-glucans from oat cultivars of Avena sativa and Avena bysantina. Journal of cereal science, 38 (1), 15-31. D. A., & Skoog, D. A. (2014). Fundamentals of analytical chemistry (9th ed.). Belmont, Calif: Thomson-Brooks/Cole.Smolyaninov, I. I. (2008). Optical microscopy beyond the diffraction limit. Standard, I. I. D. F. I.-I. (2005). Milk products and milkbased foods − Determination of fat content by the Weibull-Berntrop gravimetric method (Reference method). In. Geneva.Swagerty, D. L., D.Walling, A., & Klein, R. M. (2002). Lactose intolerance. American family physician, 65 (9), 1845-1856. Tangyu, M., Muller, J., Bolten, C. J., & Wittmann, C. (2019). Fermentation of plant-based milk alternatives for improved flavour and nutritional value. Applied microbiology and biotechnology, 103 (23), 9263-9275. Tester, R. F., & Karkalas, J. (1996). Swelling and gelatinization of oat starches. Cereal chemistry, 73 (2), 271-277. Thanavanich, C., Phuangsaijai, N., Thiraphatchotiphum, C., Theanjumpol, P., & Kittiwachana, S. (2022). Instant quantification of sugars in milk tablets using near-infrared spectroscopy and chemometric tools. Scientific reports, 12 (1), 18802-18802. M. M., Najda, A., Bains, A., Krishna, T. C., Chawla, P., Dyduch-Siemińska, M., . . . Kaushik, R. (2021). A comprehensive review on the interaction of milk protein concentrates with plant-based polyphenolics. International journal of molecular sciences, 22 (24), 13548. Tovey, E. R., & Baldo, B. A. (1989). Protein binding to nitrocellulose, nylon and PVDF membranes in immunoassays and electroblotting. Journal of biochemical and biophysical methods, 19 (2), 169-183. A., & Shanmugam, A. (2022). Study on model plant based functional beverage emulsion (non-dairy) using ultrasound – A physicochemical and functional characterization. Ultrasonics sonochemistry, 88 , 106070-106070. R., Qader, M., Jayathilake, C., Jayawardana, B. C., Liyanage, R., & Sivakanesan, R. (2020). Critical review on conventional spectroscopic α‐amylase activity detection methods: merits, demerits, and future prospects. Journal of the science of food and agriculture, 100 (7), 2836-2847. B., Guggisberg, D., Badertscher, R., Egger, L., Portmann, R., Dubois, S., . . . Rezzi, S. (2022). Comparison of nutritional composition between plant-based drinks and cow’s milk. Frontiers in nutrition (Lausanne), 9 , 988707-988707. K., Kolarich, D., Pascovici, D., Joss, J. L., Deane, E., & Packer, N. H. (2013). Characterization of N- and O-linked glycosylation changes in milk of the tammar wallaby (Macropus eugenii) over lactation. Glycoconjugate journal, 30 (5), 523-536. F., & Kemp, P. (1975). Ultrastructure of the milk fat globule membrane with and without triglyceride. Cell and tissue research, 165 , 113-127. Wooding, F. B. P. (1971). mechanism of secretion of the milk fat globule. Journal of cell science, 9 (3), 805-821. F. B. P. (1973). Formation of the milk fat globule membrane without participation of the plasmalemma. Journal of cell science, 13 (1), 221-235. P., Varka, E.-M., & Paraskevopoulou, A. (2022). Foaming properties and sensory acceptance of plant-based beverages as alternatives in the preparation of cappuccino style beverages. International Journal of Gastronomy and Food Science, 30 , 100623. K., Huang, J., Wang, D., Wan, X., & Wang, Y. (2024). Covalent polyphenols-proteins interactions in food processing: formation mechanisms, quantification methods, bioactive effects, and applications. Frontiers in Nutrition, 11 , 1371401. Zheng, B., Zhang, X., Lin, H., & McClements, D. J. (2019). Loading natural emulsions with nutraceuticals using the pH-driven method: Formation & stability of curcumin-loaded soybean oil bodies. Food & function, 10 (9), 5473-5484. Crossref Google Scholar Information & Authors Information Version history V1 Version 1 29 January 2025 Copyright This work is licensed under a Non Exclusive No Reuse License. Collection Sustainable Food Proteins Keywords analytical analytical biochemistry characterization chromatography emulsions/colloids extraction plant-based foods Authors Affiliations Tina Kayeye 0009-0003-3860-4918 [email protected] Macquarie University View all articles by this author Alison Rodger 0000-0002-7111-3024 Australian National University View all articles by this author Yuling Wang 0000-0003-3627-7397 Macquarie University View all articles by this author Anwar Sunna Macquarie University View all articles by this author John Ashton Sanitarium Health Food Company View all articles by this author Andrew Penton Sanitarium Health Food Company View all articles by this author Sarah Noyes Sanitarium Health Food Company View all articles by this author Metrics & Citations Metrics Article Usage 580 views 265 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Tina Kayeye, Alison Rodger, Yuling Wang, et al. Comparative analysis of plant and animal milks: evidence of intermolecular interactions. Authorea . 29 January 2025. DOI: https://doi.org/10.22541/au.173816988.81450559/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.173816988.81450559/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9ff5931a7cb841e2',t:'MTc3OTM4NzQxMQ=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();