Ingredients, Processing, and Fermentation: Addressing the Organoleptic Boundaries of Plant-Based Dairy Analogues

Consumer interest and research in plant-based dairy analogues has been growing in recent years because of increasingly negative implications of animal-derived products on human health, animal wellbeing, and the environment. However, plant-based dairy analogues face many challenges in mimicking the organoleptic properties of dairy products due to their undesirable off-flavours and textures. This article thus reviews fermentation as a viable pathway to developing clean-label plant-based dairy analogues with satisfactory consumer acceptability. Discussions on complementary strategies such as raw material selection and extraction technologies are also included. An overview of plant raw materials with the potential to be applied in dairy analogues is first discussed, followed by a review of the processing steps and innovative techniques required to transform these plant raw materials into functional ingredients such as plant-based aqueous extracts or flours for subsequent fermentation. Finally, the various fermentation (bacterial, yeast, and fungal) methodologies applied for the improvement of texture and other sensory qualities of plant-based dairy analogues are covered. Concerted research efforts would be required in the future to tailor and optimise the presented wide diversity of options to produce plant-based fermented dairy analogues that are both delicious and nutritionally adequate.


Introduction
Dairy remains a relevant and major agricultural product, with global milk production hitting 861 Mt in 2020 and projected to grow at 1.7% p.a. to 1020 Mt by 2030 [1]. The international demand for traditional bovine dairy remains high and stable, as shown from the minimal impact of COVID-19 on dairy production [1]. Dairy products, in particular, have played an enduring and important role in the diet of the general population, where they are consumed not only for enjoyment but also for nutritional needs and specific health benefits such as probiotic intake. Such dairy products include but are not limited to fermented foods such as yoghurt, cheese, and kefir [2]. Despite the existing popularity of conventional bovine dairy products, consumers are beginning to actively seek alternatives to them due to their potentially long-term negative impact on human health and the environment, and other ethical implications [3]. Consequently, this has led to an increased interest in plant-based dairy substitutes, which are perceived to overcome the limitations of these traditional dairy products.
Plant-based dairy substitutes are of particular interest due to their added health benefits and have been growing in popularity and market size [4]. The plant-based dairy sector is 1.
To summarise the range of raw materials used to produce dairy analogues as well as relevant physicochemical properties that support their use in such formulations; 2.
To present an overview of extraction and processing strategies to optimise these raw materials for fermentation and/or subsequent formulation of dairy analogues, with an emphasis on recent innovations (published work from 2012 to date);  3.
To discuss the impact of fermentation on the organoleptic quality of plant-based dairy analogues, with examples from both the literature (again focusing on research in the past ten years) and the market.

Common Raw Materials for Plant-Based Dairy Analogues
Various raw materials of plant origin have been utilised for the production of fermented dairy analogues and alternatives. Dairy analogues are typically made from plant materials with higher protein and/or fat content as these two components are the most essential contributors to the texture and flavour of dairy analogues [13,14]. Proteins are responsible for many physicochemical properties relevant to dairy products such as waterholding capacity, gelation, gel strength, as well as the generation of flavour precursors and/or compounds. Fats, on the other hand, affect both mechanical and sensorial properties, including mouthfeel, flavour, and flavour-carrying capacity [10,13]. Consequently, most raw materials used for dairy analogue production tend to fall into one of the four following botanical classes: legumes; grains; nuts, drupes, and seeds; tubers. Most of these materials have the added benefits of being nutrient-dense, while tubers are noted for their low cost and relative underutilisation [15,16].
The aqueous extracts of these raw materials are typically labelled as plant-based 'milks' (this term is misleading since it is not comparable to mammalian milks) [12,17]. Therefore, we will refer to these 'milks' as plant-based aqueous extracts (PBAEs) in text. With the right extraction and processing techniques, these PBAEs can be treated as dairy analogues, and further fermented and/or formulated into fermented dairy analogues such as yoghurts and cheeses [12,18]. In this section, a brief overview of the four main botanical classes of plant-based raw materials used in dairy analogue production (legumes; grains; nuts, drupes, and seeds; tubers) is provided. Selected examples of raw materials used in dairy substitute production are also included, in view of their potential to be developed into dairy analogues. The carbohydrate, fibre, total protein, and total fat contents of these raw materials are summarised in Table 1. Table 1. Carbohydrate, fibre, total protein, and total fat content of some common plant-based raw materials (all values are expressed in g/100 g dry basis).  [21] 20.1 3.0 1.6 0.1 analogues include quinoa (Chenopodium quinoa), a vitamin-rich pseudocereal containing all essential amino acids [44,48,49], and barley (Hordeum vulgare). Barley, in particular, is a carbohydrate-and fibre-rich material that is suitable for food fermentation applications. Despite the abundance of barley (accounting for more than 10% of all grains grown worldwide), less than 20% of the harvested crop is used for human consumption [15]. Collectively, grains remain underutilised as a nutritious raw material for dairy analogue production despite their mild flavour [15,47], which is a highly attractive trait for mimicking the flavour of conventional dairy; therefore, they deserve a greater amount of research interest.

Nuts, Drupes and Seeds
Nuts, drupes, and seeds are renowned for their high protein, minerals, and vitamin E content, the latter of which is known to exhibit antioxidant properties. In addition, they also contain phytosterols that reduce dietary cholesterol absorption [50,51]. Notably, their fat content ranks among the highest of all plant-based food materials. The majority of lipids found in nuts, drupes, and seeds are unsaturated fatty acids, which are beneficial to health [19]. In addition to their nutrient content, this class of raw materials has an advantage in consumer acceptability over other raw materials due to their natural nutty flavour, which is more compatible with conventional dairy flavour compared to the beany, earthy flavour of legumes and grains [52,53].
Two of the most popular dairy substitutes from this category are derived from almond (Prunus amygdalus) and coconut (Cocos nucifera). Compared to other types of PBAE, almond PBAE has a naturally creamy texture, which makes it suitable as an analogue for cow's milk and its derived products [52,54,55]. While almond PBAE is more widely consumed in North America and Europe, coconut PBAE is more frequently found in Southeast Asian countries [56,57]. Coconut PBAE has a major advantage over nut PBAEs as a raw material for dairy analogues because they do not contain typical nut allergens [57]. In general, their uniquely high fat and protein content have led to the frequent application of nuts, drupes, and seeds as raw materials in dairy analogue production [13].

Tubers
Similar to grains, tubers are another globally consumed staple crop, with the most popular tuber being the potato. These underground roots are prized for their resistance against adverse weather conditions such as drought, and they are utilised for a variety of applications [58]. Other tubers such as yam (Dioscorea spp.) and cassava (Manihot esculenta) have been growing in popularity outside of countries where they are consumed as a staple food. In addition to being rich in carbohydrates, yam contains many bioactive compounds and has been used in traditional remedies [16,59,60]. A noteworthy study by Batista et al. explored the use of yam dough in ice cream making as a substitute for cow's milk [61]. While yam is not known to be fermented for direct consumption, more studies have emerged, showing an improvement in the nutritional and sensorial aspects of this relatively underutilised crop via fermentation [16].
While various cassava snacks can be found on supermarket shelves worldwide, this crop remains most relevant in sub-Saharan Africa [21]. This is due to several properties of cassava that allow it to become a crop with high food security: the resistance of cassava roots to various pests and undesirable growing conditions; staggered harvesting as they can remain in the ground for a long time; and the highest calorie-to-cost-and-space ratio among other crops [21,60]. Additionally, a variety of fermented cascaras are consumed as staple foods in Africa, and the diversity of fermentation styles induces variations in its sensory aspects and contributes to food preservation and nutritional enhancement [58]. Despite this, cassava has not quite been explored as a dairy analogue ingredient beyond only a very few publications [62].
The tiger nut (Cyperus esculentus), a lesser-known tuber grown in Spain and Western African regions, has also been gaining increased global attention. Consumed frequently in West Africa due to its low cost and availability, the tiger nut plant is considered a weed due to its invasive nature [8,63]. However, compared to other tubers, the tiger nut has higher dietary fibre and protein content [8,51]. As such, there is increasing research interest in the traditional tiger nut 'milk' (or beverage), which is obtained by soaking, grinding, and pressing dried tiger nuts, and it has been compared to other mainstream PBAEs such as almond or soy PBAEs [8,64]. With appropriate processing and formulation strategies, high food security crops such as tubers could offer a unique advantage as cheap raw materials for dairy analogue production.

Challenges of Producing Plant-Based Dairy Analogues
Due to the vast differences in the chemical compositions between plant and dairy raw materials, mimicking both the nutrition and sensory profile of conventional dairy products continues to be a key challenge in the creation of dairy analogues. Additionally, though soy and other legumes have been used extensively as milk alternatives or analogues, their naturally 'beany' odour is perceived as an off-flavour in the context of dairy profiles [4]. These raw materials may also contain antinutrients such as inositol phosphate, which reduces the nutritional quality of the plant-based products by impeding the absorption of nutrients such as minerals [44,45].
The shortfalls of plant-based dairy analogues versus their dairy counterparts have often been addressed by the addition of stabilisers, fillers, nutrients, and other processing aids [32,45,65]. For example, oils are commonly added for flavour and texture purposes and lecithin for emulsion stabilisation [4]. Plant-based dairy analogues also tend to be fortified with vitamins (e.g., A and D) and minerals (e.g., calcium) to nutritionally resemble bovine milk [4]. However, this may decrease their appeal to consumers due to increasing demands for clean-label products and the known flavour and nutritional issues with ultra-processed plant-based foods [10,66]. Consequently, research has turned to the use of fermentation to narrow the gap between dairy analogues and their benchmarks without excessive additive use. Numerous studies have demonstrated the positive impact of fermentation on the nutritional and organoleptic qualities of plant-based dairy analogues, including but not limited to the degradation of antinutrients, probiotic function, textural improvement, and off-flavour reduction [4,37,47,65]. These will be covered in detail in Section 4 of this review.
Nonetheless, the application of appropriate raw material extraction and processing strategies is warranted to complement and amplify the positive impacts of fermentation on the organoleptic quality of plant-based dairy analogues. These techniques comprise a range of mechanical, chemical, biological, and novel processing methods, with a common aim of obtaining a matrix with the best functional properties for the subsequent production of dairy analogues. The next section of our review provides insights into recent studies on these extraction and processing strategies, with a key focus on organoleptic improvement.

Extraction and Functionalisation of Ingredients from Plant Raw Materials
Aside from the selection of raw materials with suitable flavour and nutritional properties, the extraction of such plant-based materials to generate suitable ingredients is also critical to the development of a satisfactory plant-based dairy analogue. Extraction processes have a profound effect on the composition of the raw material, which then determines its behaviour during subsequent product development stages [7]. As discussed in Section 2, the extraction and characteristics of fat and protein are critical considerations in the production of dairy analogues due to their effect on the ingredient's functional properties and flavour [67]. In particular, extraction treatments as well as the relevant pre-and post-extraction treatments (depicted in Figure 1) can lead to varied ingredient microextractions and protein conformations. These can affect solubility, water absorption capacity, gelation, and emulsion stability, all of which ultimately modify the end product texture [67]. Based on the organoleptic limitations of plant-based raw materials (Section 2), considerations regarding the generation of flavours and reduction or elimination of offflavours are also warranted [17]. These organoleptic qualities are the main determinants of product acceptability and, hence, are the focus of this discussion, though we will provide Foods 2022, 11, 875 7 of 40 a brief discussion on the effect of these extraction techniques on the recovery of some chemicals of health interest [68]. A summary of these studies is provided in Table 2. end product texture [67]. Based on the organoleptic limitations of plant-based raw materials (Section 2), considerations regarding the generation of flavours and reduction or elimination of off-flavours are also warranted [17]. These organoleptic qualities are the main determinants of product acceptability and, hence, are the focus of this discussion, though we will provide a brief discussion on the effect of these extraction techniques on the recovery of some chemicals of health interest [68]. A summary of these studies is provided in Table 2.  In the process of creating a dairy analogue, two main ingredient classes derived from plants are considered: solid flour extracts and PBAEs. PBAEs tend to be the ingredient of choice for the development of dairy analogues and will be reviewed in greater depth. It is also worth noting that different plant raw materials can exhibit various trends upon treatment, and these differences are discussed to provide an overview of the diversity of manipulations and their effects on ingredient quality [9].

Conventional Mechanical Operations
For most studies, the mechanical extraction of plant raw material is applied. Mechanical treatments allow for standardisation of the process and are effective in dispersing the raw material for ingredient extraction. In addition, they are low cost, scalable, and have a low technological barrier [7,17]. Along with these mechanical operations, thermal treatments may be conducted at several points to improve extraction yield or ingredient properties.

Pre-Treatments
Roasting-Prior to milling, a variety of pre-treatments may be conducted. Thermal treatment with roasting has been applied to specific raw materials ( Table 2) such as legumes and peanuts to enhance the flavour and aroma of their extracted flours and PBAEs [17,67,69,70] and was also noted to reduce LOX-generated off-flavours in sesame [70]. The sensorial acceptability for such thermally generated flavours ('roast' notes) in dairy analogues, however, requires further study. Roasting also contributes to protein denaturation, which alters the ingredient's functionality [7]. It reduced protein solubility for sesame PBAEs [70] but conversely increased protein solubility and emulsion stability for peanut PBAEs [69]. More studies on the impact of roasting parameters and their variation with raw materials are needed for a better understanding of its influence on protein functionality.
Dehulling-Dehulling is suitable for isolating desired elements from raw materials that possess numerous components (e.g., legumes and grains) [23] (Table 2). Both dry-dehulling and wet-dehulling may be conducted, although there are few studies observing the impact of this selection on the functionality and flavour of plant-based dairy analogues. Dehulling removes dietary fibres as demonstrated in several legume flours [72], which may result in smoother, less gritty, and more pleasant textures in dairy analogues. The resulting concentration of the endosperm may also remove off-flavours and antinutrients for a variety of legumes [72,100]. However, the impact of dehulling depends on downstream processes such as soaking. Ma et al. demonstrated that dehulled peas potentially released more 'small molecules' during soaking, which resulted in lower amounts of the off-odorant 2-methoxy-3-isopropyl-(5/6)-methyl pyrazine and albumin (which did not negatively impact texture in this study) in the obtained pea PBAE [71]. It is thus evident that the effects of pre-extraction treatments can be interdependent.
Soaking and blanching-The raw plant material may be either soaked in cold water or blanched in hot water to soften it and remove undesirable water-soluble components [9,17,75] (Table 2). While the additional heat from blanching can result in more efficient removal or inactivation of undesirable off-flavours and antinutrients compared to soaking, it can also affect the functionality, flavour, and nutritional value of the final ingredient [12]. For example, flours from boiled and roasted pulses (seeds of legumes) displayed 2-3 times higher water absorption capacities and higher gelation rates than flours from raw pulses [67]. Heat treatment from blanching was also found to inactivate undesirable endogenous enzymes to further improve flavour and nutrition [9], such as LOX that produce off-flavours in soy and peanut PBAEs, and trypsin or other protease inhibitors that restrict protein digestion [7,67,101]. However, blanching can also result in the loss of desirable nutrients such as proteins, choline, and folate, although the extent of such losses is dependent on the raw material and temperature used [67,101]. While the impact on functionality and nutrient composition can be quite varied, overall, soaking, and especially blanching, are effective methods for off-flavour removal.

Extraction and Separation
Mechanical grinding is then performed on the untreated or pre-treated plant tissue [9]. For flour ingredients, dry-milling is usually first conducted, and the flour would later be reconstituted or further extracted to create the final product [17]. For PBAE ingredients, aqueous extraction is required, and the raw material is typically ground into a slurry by wet-milling to release soluble or finely suspended materials [7,17]. Both the resulting flours and PBAEs possess a non-homogenous particle size distribution that may require further size standardisation or reduction for texture and stability [64]. Superfine pulverisation technologies have been gaining attention for plant flour ingredients such as colloid-milling, jet-milling, and ball-milling [23,102]. Related to this point, homogenisation for PBAE ingredients is discussed in Section 3.3.1.
Conventional milling processes can be innovated by tailoring parameters such as oxygen availability and temperature, which affect the composition and organoleptic qualities of the extracted plant ingredient. Kaharso et al. recently investigated anaerobic wet-milling of soy with oxygen-free water and found that it significantly reduced the formation of lipid oxidation products and off-odorants (e.g., alcohols and aldehydes) in the resulting soy PBAE [74]. Thermal treatments are also commonly applied for wet-milling (e.g., cooking the slurry) to increase extractability [7,17]. However, the impact of thermal treatment is not straightforward. For example, high temperature cooking increases nutrient solubility and recovery, but extreme temperatures can denature plant proteins and decrease their yield and/or alter functionality [7,103]. Thermal treatment during or directly after milling makes a significant contribution towards the deactivation of endogenous enzymes to reduce offflavour production and antinutrient content, especially for soy [103]. High temperature treatment can also increase oil extractability [7,17] and affect starch gelatinisation [7,13]. These effects on downstream processing and organoleptic qualities on oil-rich nuts or starch-rich grains can be positive, such as the generation of desirable flavours and textures (e.g., in yoghurts), or negative, where extra processing steps may be required.
Finally, separation of the unwanted material (typically coarse particles in the PBAE) or concentration of desirable components (e.g., proteins) occurs by decanting, gravity, centrifugation, or (ultra)filtration [9,104] to remove excess lipid materials and prevent the coalescence of oil bodies and phase separation, which ensures product stability and a consistent lipid proportion in the ingredient [54,65]. If necessary, a final drying step (e.g., spray-drying) may also occur for easier transportation or incorporation of the completed ingredient.

Chemical and Biological Aids
Although mechanical methods form the backbone of raw material isolation and functionalisation strategies for flour or PBAE creation [17], there has been increased incorporation of chemical and biological techniques to improve the resulting ingredient quality for flavour or fermentation purposes [9]. Off-flavours remain a major concern for plant-based ingredients as these significantly hinder their applicability to bovine-milk-based products, especially fermented products such as yoghurt and cheese, as consumers are sensitive to the typically 'grassy' or 'earthy' off-flavours [81]. A variety of chemical and biological techniques have thus been applied to strategically deodorise the ingredients yielded from plant raw materials or to functionalise other components.

pH Treatment and Other Chemical Extraction Techniques
pH alteration-The pH of the soaking or extraction environment has commonly been altered to facilitate protein extractability, especially to manipulate the solubility of proteins based on their isoelectric points, which may be acidic or alkaline depending on the raw material [17,70,75] (Table 2). In cases where powdered protein isolates are the target ingredient, pH alteration could be utilised for isoelectric precipitation to induce protein aggregation for subsequent extraction [4,13]. More interestingly, pH alterations can also be applied to reduce off-flavour formation. Ahmadian-Kouchaksaraei et al. reported higher LOX activity of sesame seeds in acidic conditions (pH 5) [70], and it is a common industrial practice to alkalinise soy or peanut PBAEs with sodium bicarbonate to reduce LOX activity and lipid oxidation off-flavours [70,105]. pH also affects protein solubility and functionality [75] and was manipulated by Ma et al. to alter gel hardness for pea yoghurt, which resulted in softer and more sensorially acceptable yoghurts [71].
Chemical and physical deodorisation-Such techniques are largely used for plant protein isolates or plant flour ingredients, and despite their promise for off-flavour reduction, have not been directly assessed for use in plant-based dairy analogues [76][77][78] (Table 2). Wang et al. demonstrated that alcohol washing of pea protein flour resulted in deodorisation and the removal of off-flavours, though high alcohol washes led to reduced emulsion stability and solubility of the pea flours [78]. Sorptive techniques are less common but also effective for off-flavour removal. For example, polystyrene and zeolite-based adsorbents were applied to remove the majority (60-70%) of hexanal from soy protein isolate [31]. Most commonly, vacuum treatment at high temperatures has been applied to soy and peanut PBAEs to indiscriminately strip it of its characteristic aromas [105,106] that would be unpleasant in the context of dairy products. Distillation techniques (physical deodorisation) such as supercritical carbon dioxide extraction have been effective in removing off-flavours from pea flours [76] (Table 2). As these chemical methods (along with solvent and sorptive extractions) effectively strip the raw material of most odorants, the aroma would have to be reintroduced to the product through other ingredients or fermentation strategies.

Enzymatic Treatments
The variety of commercially available food-grade enzymes has vastly improved the quality of plant-based ingredients. Plant materials differ from bovine milk due to the presence of fibre as well as large oligomeric proteins, which results in gritty textures [81] and reduced emulsion stability, leading to undesirable mouthfeel in dairy analogues [7]. The majority of the enzymes applied to plant-based dairy analogues are used for the hydrolysis of macromolecules to reduce particle size and improve solubility and mouthfeel [81]. Li et al. demonstrated that papain treatment in soy cheese was able to hydrolyse proteins and yield a more homogenous protein network with improved textural and hedonistic qualities [81]. Luana et al. also demonstrated that enzymatic (Depol 740 L and Grindamyl 1000) treatment of an oat yoghurt beverage during fermentation led to improved aroma and taste characteristics as well as reduced fermentation latency, possibly due to its release of free sugars and other fermentation substrates [53]. As such, enzymatic treatment is especially complementary to the fermentation process in improving end product organoleptics.
Another major application for enzymes in PBAE ingredients is starch liquefaction (as mentioned in Section 3.1.2), typically with α/β-amylases, to reduce the viscosity and improve the fluidity and emulsion stability of the PBAE [14,19], which may also be beneficial for releasing free sugars for subsequent fermentation. While enzymes are a powerful tool to improve the functionality of plant-based dairy ingredients and analogues, they are highly specific and costly. Where the use of multiple enzymes is warranted, fermentation may prove to be a cheaper and more sustainable alternative.

Sprouting and Germination
Another cost-effective alternative to enzymatic treatment is the natural germination of the plant raw material. During germination, proteases and amylases among other enzymes are activated, which can significantly alter the composition and functional properties of the plant material [4,82]. Decreases in the water-holding capacity of soy yoghurts were observed with germination (and degree of germination), likely due to starch hydrolysis, which altered the gelation properties of soy PBAE and yoghurt, resulting in more sensorially acceptable textures. Germination has also been reported to improve the flavour characteristics and sensory acceptance of rice and soy yoghurts [83][84][85] (although the opposite was reported for sprouted tiger nut yoghurt [82]), due in part to a reduction in LOX activity with germination, which is a trend observed in PBAEs [24]. Germination-derived increases in free sugars and amino acids may serve as valuable precursors for the formation of pleasant aroma compounds during fermentation [85]. This also explains the observation that germinated plant material improves the growth kinetics of starter cultures during fermentation, which synergises with the application of fermentation for the organoleptic improvement of plant-based dairy analogues [83].
Aside from texture and flavour benefits, germination was found to improve the nutritional properties of fermented plant-based dairy analogues by decreasing the antinutrient content or increasing the amount of γ-aminobutyric acid owing to increased glutamic acid decarboxylase activity [82,84]. For studies that focused on the extraction of PBAEs, germination was also observed to increase the ingredient's antioxidant activity and decrease antinutritive content (e.g., saponins, phytate, trypsin inhibitors) depending on the germination duration (Table 2) [4,24,107,108]. As germinated plant ingredients are viewed by consumers to be nutritionally superior and natural, germination appears to be extremely viable for improving ingredient and fermented product quality.

Enhanced Functionality with Innovative Processing
For plant-based ingredients to be made into dairy analogues, the main challenges are the texture, off-flavour, and antinutrient content [47,109]. When it comes to texture and stability, PBAEs possess thermodynamically unstable and polydisperse particle distributions due to their diversity of particles (e.g., oil droplets, native protein aggregates, polysaccharides, and cell fragments) [4,110]. This makes PBAEs especially prone to sedimentation, creaming, or syneresis over storage [4,109,110], which complicates their application as ingredients for dairy analogues. Most conventional mechanical or chemical processing methods do not address this aspect of the plant material microstructure. Hence, greater interest has arisen in innovative processing technologies, and some achievements and limitations are discussed here.

Homogenisation by HPH and Ultra-HPH (UHPH)
Homogenisation has a huge impact on the extracted PBAEs' microstructure [111] and is also often applied to bovine milk for texture and stability improvement through the reduction and standardisation of the size of fat globules [9,54]. Due to the different microstructures and constituents of plant-based ingredients, conventional homogenisation parameters for bovine milk are unlikely to perform equivalently well in PBAEs. The particle heterogeneity of PBAEs demands more aggressive processing parameters than those used in bovine milk (10-25 MPa) [109], and some studies have investigated the potential benefits of HPH and ultra-HPH (UHPH) to improve the stability and organoleptics of PBAE ingredients.
HPH and UHPH could assist in achieving artificial globules that resemble bovine milk viscosity, stability, and mouthfeel for organoleptic purposes [9] as well as modify protein conformation and functionality (e.g., emulsifying or foaming properties) for texture improvements. This particularly benefits yoghurt analogue production by creating PBAE ingredients with improved textural properties. Ferragut et al. demonstrated in soy-based yoghurts that UHPH (200-300 MPa) soy PBAE resulted in preferable textures compared to thermally treated soy PBAE. UHPH-treatment was found to increase the onset of gelation and decrease aggregation rate and gel network density, resulting in improved mechanical properties such as higher firmness, water-holding capacities, and more compact network structures in the soy yoghurt product [89,112,113]. Demirkesen et al. also demonstrated that the microfluidic treatment of hazelnuts (135 MPa) to create a whole hazelnut PBAE resulted in improved hazelnut yoghurt texture, which was closer to that of conventional yoghurt [88]. Generally, the application of HPH techniques to PBAEs resulted in highly stable emulsions with viscosities and mouthfeels similar to bovine milk (Table 2) [91], which makes HPH an attractive processing strategy in dairy analogue production [114].
Although less well studied, HPH could also improve the flavour of plant-based ingredients. Poliseli-Scopel et al. and Pérez-González et al. both studied changes in the volatile composition of soy and almond PBAEs, respectively, with UHPH compared to other processing methods [115,116]. They noted that UHPH-treated PBAEs tended to attain improved sensory characteristics, particularly with lower lipid oxidation markers, which may indicate lower off-flavour formation by LOX due to inactivation by the high shear forces ( Table 2).

Innovative Non-Thermal Technologies
Heating in the form of pasteurisation and ultra-high temperature (UHT) treatment is commonly applied for microbial inactivation in PBAE ingredients, though this can lead to the formation of thermal off-flavours in plant-based dairy analogues [68]. While many innovative non-thermal techniques have been assessed for microbial inactivation purposes [68], they may also play additional roles for texture and flavour stability. As such, modern techniques (e.g., ultrasonication, high hydrostatic pressure (HHP), and pulsed electric field (PEF) treatment) that have been used in plant-based dairy applications were evaluated.
Ultrasonication-Aside from HPH, ultrasonication has also been explored for the reduction and standardisation of particle sizes and for protein modification. Ultrasonication was found to reduce particle size and distribution in coconut PBAEs, resulting in greater stability and increased fluidity [93,94,117]. Depending on the ultrasound parameters, a variety of protein modifications may also result in changes to the ingredient's textural functionality, as reviewed by Gharibzahedi et al. for legume proteins [118]. While ultrasonication appears to have large benefits in terms of ingredient texture, it tends to promote destructive oxidative processes [118], which may affect the flavour of plant-based dairy analogues.
HHP-For the creation of dairy analogues, HHP is of interest due to its impact on protein functionality. HHP was found to induce gelation with plant protein materials, which may be applicable for certain products such as yoghurts [96] and was successfully applied to increase water-holding capacity and reduce syneresis in soy yoghurts [95]. For materials where off-flavour production is enzymatic in nature, HHP may also improve sensory characteristics by enzyme inactivation, as demonstrated by its ability to reduce LOX activity in soy yoghurts and PBAEs resulting in fewer lipid oxidation off-odorants [95]. HHP has also been shown to reduce the allergenic characteristics of plant proteins from the raw material, which is beneficial for consumer health.
PEF-PEF involves placing the food product between two electrodes and subjecting them to short pulses at high voltages for a relatively short processing time (compared to thermal treatment). Thus far, PEF has been applied to PBAE ingredients and has mainly appeared to be effective for reducing PBAE particle size and increasing colloidal stability [119]. PEF may be similar to HHP [120] in that it deactivates enzymes that result in off-odour formation ( Table 2) [98], showing potential for its use for dairy analogues.

Blending Plant Materials and Additives
Ultimately, each raw material faces unique drawbacks for use in plant-based dairy analogues due to a mix of organoleptic and nutritional disadvantages that may be extremely challenging to evade even with sophisticated processing methods. Blending plant raw materials at different ratios has thus been of great interest to improve the flavour profiles, textures, and nutritional properties of dairy analogues [32]. Coda et al. blended cereal and soy flours with concentrated grape must to create yoghurt beverage analogues, and they found that mixtures of rice and barley or emmer flours resulted in improved organoleptic and nutritional properties compared to pure rice flour ferments [121]. Adejuyitan et al. found that a 50:50 soy/coconut PBAE-based cheese substitute yielded a higher hedonistic rating than a 100% soy version [122]. Additionally, Oyeyinka et al. demonstrated that a cheese analogue containing a 2:3 ratio of cashew to soy PBAE achieved the highest protein content and hedonistic rating for flavour [123]. While these studies showed that blending raw materials could result in improved hedonistic ratings, the mechanism behind this is unclear, although Short et al. suggests that one potential reason could be the masking of off-flavours, e.g., the beany note in soy [32]. Blending raw materials is also of great nutritional interest, especially for the amino acid profile of the ingredients. As discussed in Section 1, plant materials, unlike bovine dairy, have lesser amino acids and comparatively poorer digestibility [4]. Hence, purposeful blending of plant materials could help to achieve a more complete amino acid profile without fortification. In summary, blending different raw materials could result in an ingredient with improved properties for subsequent formulation or fermentation into a plant-based dairy analogue with improved sensorial and nutritional properties.

Fermentation as a Strategy to Improve Organoleptic Properties of Plant-Based Dairy Analogues
Collectively, the selection of appropriate raw materials, extraction, and processing strategies complements the use of fermentation in creating a plant-based dairy analogue with an authentic, dairy-like sensory profile. Fermentation can play a defining role in generating dairy-like flavour and textural attributes in plant-based dairy analogues. These include the modification of sensory characteristics such as acid production, masking or eliminating native off-flavours in plant materials, and secreting exopolysaccharides (EPS) that thicken the plant matrix to emulate the creamy texture of dairy products [13,124]. Fermentation has been shown to improve the nutritional content of plant-based dairy analogues by increasing the bioavailability of nutrients, reducing antinutritive components and/or allergens, as well as additional probiotic functions [4,125]. It also improves the safety and shelf life of these analogues through acidification, the generation of antimicrobial compounds, and competition with undesirable microorganisms [125][126][127].
There exists a multitude of starter cultures available to the researcher or manufacturer seeking to apply fermentation to the development of a plant-based dairy analogue. Table 3 summarises some examples of fermented plant-based dairy analogues and starter cultures that are available in the market. In the interests of enhancing the organoleptic properties of a dairy analogue, many studies and commercial products utilise starter cultures associated with traditional fermented dairy products. A brief overview of such cultures is given in the following sections before the provision of specific examples in different product categories (fermented cream products, yoghurt, cheese, kefir). Interestingly, the complexity of mimicking an authentic dairy profile-be it flavour, texture, or appearance-wise-may lead to the use of strains not usually associated with dairy fermentations [128].      Of the array of microorganisms involved in food fermentation, lactic acid bacteria (LAB) are an indispensable group when it comes to dairy fermentation. They are a phylogenetically heterogenous group comprising Gram-positive, non-motile bacteria including the genera Lactobacillus (L.), Lactococcus (Lc.), Leuconostoc (Leu.), and Bifidobacterium (B.), to name a few. Different members of the LAB family may perform homolactic or heterolactic fermentation based on hexose catabolism and other environmental factors. Homolactic fermentation is common in LAB associated with meat and dairy fermentations where acidification is the primary function of fermentation, and the major end product is lactic acid. Heterolactic fermentation, on the other hand, tends to be more prevalent in plant-based fermentations compared to meat and dairy, where it generates significant amounts of acetic acid, ethanol, and CO 2 along with lactic acid. Other metabolic processes relevant to the sensory aspects of fermented dairy include citrate utilisation that generates acetic acid, lactic acid, diacetyl, acetoin, and CO 2 [129,130].
There have been several taxonomic changes to the lactobacillus group of late. However, they have yet to be validly published (considered as accepted, now published in [131] and being adopted) under the rules of the International Code of Nomenclature of Bacteria [131,132]. For the purposes of this review, we will continue using the old nomenclature of Lactobacillus for the following species: Lactiplantibacillus plantarum, Lactiplantibacillus pentosus, Lacticaseibacillus casei, Lacticaseibacillus paracasei, Lacticaseibacillus rhamnosus, Limosilactobacillus fermentum, Limosilactobacillus pontis, Limosilactobacillus reuteri, Latilactobacillus curvatus, Fructilactobacillus sanfrancisensis, Levilactobacillus brevis, Liquorilactobacillus nagelii, Lentilactobacillus hilgardii, Lentilactobacillus buchneri, and Lentilactobacillus kefiri.
In dairy fermentations, LAB play a crucial role in flavour formation, textural modification, and food preservation. The application of probiotic strains confers additional health benefits to the consumer [127,129,133]. Their safety and ubiquity are well-documented, making them a popular choice in fermentation studies, even if the raw material in question is not a native environment for LAB growth [4,134,135]. LAB often assume the role of starter cultures in dairy fermentations, e.g., St. thermophilus and L. bulgaricus in yoghurt, and L. kefiranofaciens and L. kefiri in milk kefir [133,136]. They are also used as adjunct cultures for organoleptic improvement, particularly in cheeses. These non-starter LAB (NSLAB) include strains such as L. casei, Leu. mesenteroides, Leu. dextranicum, and L. plantarum [137].

Yeasts and Filamentous Fungi
In contrast, yeasts and fungi are seldom used as starter cultures in dairy fermentations. Many strains are associated with surface-and mould-ripened cheeses [138,139]. Yeast and fungi play several distinct roles in cheese fermentation. Some species ferment lactose, such as Kluyveromyces lactis, K. marxianus (Candida kefyr), and more rarely, Saccharomyces cerevisiae, while others assimilate lactose and/or galactose, e.g., Debaryomyces hansenii, Geotrichum candidum, and Penicillium camemberti. Additionally, G. candidum, P. camemberti, and some strains of D. hansenii assimilate lactate, raising the pH of the cheese matrix and hastening the process of ripening. Several species are especially prized for their strong proteolytic and lipolytic activities, including G. candidum, Yarrowia lipolytica, and several Penicillium spp. Their broad enzymatic activity liberates free amino acids and fatty acids in the matrix, consequently imparting the cheesy, umami, and bitter notes we associate with cheese [137,138,140,141]. Mould-ripened cheeses such as Brie, Camembert, and blue-veined cheeses owe their distinctive appearances and flavour to fungal adjunct cultures.
Spontaneous fermentations are responsible for the birth of fermented milk drinks, including kefir, koumiss, gariss, and chal. They are generally recognised to be the product of mixed culture, yeast-LAB fermentations, and their primary differences lie in microbiota composition and milk source [142][143][144]. Though the production of different fermented milks is traditionally restricted to specific regions and cultures, there is increasing interest in popularising such beverages for their purported health benefits [145,146]. Yeasts that ferment or assimilate lactose predominate in fermented milk matrices, including C. kefyr, K. marxianus, and K. lactis. All other associated yeast species generally metabolise lactate, usually under aerobic conditions, such as Issatchenkia orientalis (Candida krusei), Y. lipolytica, and Saccharomyces unisporus. Reminiscent of cheese fermentation, lactose-fermenting yeasts ferment the residual lactose in the medium and/or raise the pH through lactate assimilation. At the same time, they produce substantial quantities of ethanol and carbon dioxide, giving some fermented milks their signature fizzy and alcoholic flavour [144,[147][148][149].

Fermented Plant-Based Dairy Analogues
The following subsections provide selected examples of studies performed in the last ten years (2012 to date) pertaining to the organoleptic improvement of plant-based dairy analogues via fermentation. These include the enhancement of flavour, texture, appearance, and/or other sensory characteristics to better mimic a defined dairy benchmark of the plant-based analogue. As such, several types of studies are excluded from this review, including but not limited to: (1) studies simply assessing probiotic viability in plant-based raw materials; (2) the application of classic dairy fermentation cultures and protocols (yoghurt, kefir, etc.) in plant matrices without evaluation of their organoleptic properties or comparison against a dairy benchmark; (3) nutritional studies on fermented, plant-based dairy analogues. A brief overview of the studies included is summarised in Table 4. Table 4. Selected studies in the past ten years (2012 to date) applying fermentation as a strategy to improve the organoleptic properties of plant-based dairy analogues.   Off-flavour reduction in cheese analogue  ). 2 The complete scientific name of each strain is always provided where possible, e.g., St. thermophilus, L. bulgaricus. 3 Where a commercial starter culture is used, the product name is listed, followed by the specific cultures in parenthesis. * A probiotic L. rhamnosus strain was inoculated for viability tests during prolonged storage but was not deemed to have affected the organoleptic quality of the product.
In literature, the fermentation of plant-based milk analogues is often designed with nutritional improvement in mind; many studies also evaluate probiotic viability [4,65]. Notably, a study by Tangyu et al. screened a number of microorganisms for their potential in enhancing the nutritional and sensory profile of chickpea-based milk analogues [37]. A significant decrease in off-flavour aldehydes was observed, accompanied by an increase in sweet, fruity, and creamy notes, especially with citrate supplementation. Commercially available plant-based milk analogues, however, generally rely on processing techniques and/or formulation strategies to mask unpleasant notes and resemble the organoleptic profile of dairy milk [4,65]. It is plausible that fermentation is rarely practiced for the organoleptic improvement of commercial milk analogues because it is less economically feasible than other processing strategies, or that it may result in acidification, which could curdle plant proteins and negatively affect the fluidity of the milk analogue [13].

Fermented Cream Products
Butter (lactic butter), lactic or traditional buttermilk, and sour cream are all products of cream fermentation. To make lactic butter, cream-typically from cow's milk-containing at least 40% fat is fermented with pure or mixed LAB cultures of Lactococcus lactis and cremoris, Lc. lactis biovar. diacetylactis, and Leu. cremoris or other diacetyl-producing LAB strains before churning; the liquid expelled during this process is known as buttermilk. Nowadays, it is more common for buttermilk to be made by directly fermenting semiskimmed milk with the same strains used in butter fermentation. A similar process applies to the manufacture of sour cream, though no churning is involved [127].
At the time of writing, we had only identified one study on fermented plant-based cream products that fell within the scope of our review. Madsen et al. explored the possibility of producing a buttermilk koldskål (Danish cold buttermilk soup) analogue by fermenting tiger nut extract with plant-isolated LAB and conventional yoghurt strains. The results indicated that fermentation by Leu. mesenteroides (isolated from gooseberries) mimicked the acidity of commercial koldskål and generated an aroma profile similar to sweet fermented milk. Xanthan gum was, however, needed to improve the body and stability of the analogue, as fermentation alone could not reproduce the viscosity of traditional koldskål [63].
In contrast, there are numerous plant-based analogues of fermented cream products on the market ( Table 3), consisting of butters, sour creams, and sour cream-based dips. Cashew, almond, coconut, and oat PBAEs often serve as the base ingredient, along with plant-based oils. Starches, gums, and plant-derived lecithin are common features in the ingredient list, likely as textural aids. Most products did not disclose the specific cultures used for fermentation. Of note, wildbrine indicates the use of lactobacilli in its butter alternatives, while Forager Project cultures its sour cream analogue with LAB that are commonly employed in yoghurt fermentation (Table 3).

Yoghurt
Yoghurt is arguably the most popular and diverse fermented dairy analogue. Set, stirred, drinkable, flavoured-the seemingly endless variations of yoghurt products on the market belie the humble makeup of its microbiota. Traditionally, yoghurt is fermented from whole cow's milk with only two cultures-St. thermophilus and L. bulgaricus. Though yoghurt may now be made with additional cultures such as Lactobacillus acidophilus, L. casei, and Bifidobacterium spp., its pool of starter cultures is comparatively small compared to cheeses and kefirs [127,136,138]. During yoghurt fermentation, lactose is fermented to form lactic acid, which gives yoghurt its signature tangy flavour and induces the acid gelation of casein, a major protein in milk. This forms the firm, viscous, and cohesive gel characteristic of dairy yoghurts [22,127]. The absence of casein in plant matrices as well as dairy-incompatible, native plant flavours are thus key hurdles towards the development of a stable, palatable, and dairy-like yoghurt analogue [22,158,161].
Commercial dairy yoghurt cultures and probiotics are typically applied in the fermentation of plant-based yoghurt analogues (Tables 3 and 4). They have been shown to induce the formation of yoghurt-like gels, impart sweet, creamy aromas reminiscent of cow's milk yoghurt, and in some situations, mask or reduce the perception of off-flavours such as beany notes [27,82,114,150,151,154,155]. Meanwhile, studies on suspensions of lupin protein isolate [158] and a rice-chickpea-lentil mixture [40] indicated that several novel, non-yoghurt strains are also capable of producing a yoghurt-like odour and/or texture. Like Madsen et al., these studies demonstrated that the optimal fermentation protocols for dairy analogue production may not include traditional dairy starters, which highlighted the importance of tailoring fermentation cultures and conditions to each plant matrix. This was observed by Luana et al. as well, where fermentation with L. plantarum LP09 (a non-dairy strain) reduced the perception of earthy, cereal notes while boosting the intensity of acid and dairy-like notes in a fermented, oat-based yoghurt drink [53]. Similarly, soy PBAE fermented with bifidobacteria produced higher levels of acetaldehyde, an important aroma compound of dairy yoghurt, compared to conventional yoghurt starters [33].
Plant-based yoghurts produce weaker gels than their dairy counterparts and are highly prone to syneresis, regardless of the type of raw material used [36,57,153,156].
While the discussed processing strategies (Section 3) are helpful, researchers have been developing fermentation protocols which can simultaneously achieve optimal flavour and texture in the final yoghurt analogue. This has led to the use of EPS-producing LAB, which synthesise and excrete largely taste-neutral EPSs that exhibit hydrocolloidal behaviour and bind water efficiently, improving the growth medium's rheological and textural properties [26,48,49,158]. EPS-producing LAB are often isolated from yoghurt, kefir, and sourdough starters, though some are associated with beer spoilage. Beyond serving as a source of microbial-derived EPSs for food applications, EPS-producing LAB have been used in fermentations to improve the texture of a wide range of products [162,163].
Fermentation with EPS-producers has been explored in plant-based yoghurt analogues made from soy PBAE, lupin protein isolate, and quinoa flour [26,48,49,158]. Li et al. compared the effects of soy yoghurt fermentation using EPS-producing LAB versus a commercial yoghurt starter culture. As well as producing the highest apparent viscosity, the content of beany flavour compounds (hexanal, 2-pentylfuran, 2-pentanone) decreased in soy yoghurt after fermentation with an EPS-producing L. plantarum strain, while 3hydroxy-2-butanone, a characteristic flavour compound of fermented dairy milk, increased to detectable levels [26]. Similarly, high viscosity was observed in the fermentation of quinoa-based yoghurt with EPS-producing Weissella cibaria and W. confusa, the latter of which also imparted sweet and dairy-like acidity [48,49]. A survey of 30 different LAB strains in yoghurts made from lupin protein isolates-which possess weak gelling abilityrevealed that EPS-producers (L. plantarum, Pe. pentosaceus, L. brevis) were the most effective in emulating both the aroma and texture of dairy yoghurt [49,158].
Fermented, plant-based yoghurt analogues on the market are almost exclusively made from nuts, drupes, and seeds (cashew, almond, coconut). Save for Alpro's oat yoghurts, which feature the cheese starters Lc. lactis and Lc. cremoris, most brands use traditional yoghurt cultures and their associated probiotic lactobacilli (Table 3). Starch, gums, and/or pectins appear in every ingredient list. EPS-producing LAB do not appear to be used for commercially available yoghurt analogues; if they are, they remain unspecified. A variety of starter cultures are also available for at-home fermentation of plant-based yoghurts; most include dairy yoghurt starter cultures and probiotics (Table 3).

Cheese
Variations in milk source, starter and adjunct cultures, fermentation conditions, ripening operations, and other cheesemaking processes are responsible for the close to 1500 cheese varieties worldwide [13,137,138]. Despite significant differences in flavour, texture, and appearance among cheese varieties, cheesemaking generally results in a common aim-the production of a viscoelastic curd as a result of the agglomeration of caseins in fluid milk [13,164]. Starter cultures in cheese, namely LAB, acidify the matrix via lactose fermentation to promote cheese curd formation. They can be mesophilic (Lc. lactis and Lc. cremoris) or thermophilic (St. thermophilus) in nature.
Adjunct cultures flourish in the cheese environment during the later ripening stage, where they metabolise various substrates to produce gas, colour, or characteristic flavour compounds in cheese [137,164].
Due to their larger molecular size and substantial structural differences, plant proteins do not exhibit similar aggregation or gelation behaviours as casein micelles [10,11]. While the inclusion of fillers and stabilisers are helpful, recent studies have highlighted the significant influence of extraction and processing parameters on the texture of cheese analogues (Table 2) [18,159,165]. Though fermentation and its resultant acidification have been used to curdle PBAEs [29,81,160], it is rarely explored as a specific strategy for textural improvement.
Fermentation by LAB and commercial cheese cultures was, however, observed to improve gel hardness and reduce syneresis in soy-based petit-suisse (fresh cheese) analogues [29] and pea protein isolate-olive oil emulsions [41]. It is important to note that product formulation, including the use of gums and vegetable fats, played a significant role in the textural stability of the aforementioned gels.
Other researchers have turned to the use of fungi instead, particularly G. candidum, an adjunct culture responsible for the velvety appearance of Camembert and Reblochon, among other cheeses. As well as raising the pH of the cheese matrix via lactate metabolism, G. candidum is prized for its strong proteolytic and lipolytic activity and its ability to produce volatile sulphur compounds (VSCs) [137]. Łopusiewicz et al. [152] studied the production of a Camembert analogue by fermenting flaxseed oil cake with LAB starters (Lc. lactis, Lc. Cremoris, and St. thermophilus), P. camemberti, and/or G. candidum. The inclusion of G. candidum resulted in a significantly lower hardness and chewiness in the Camembert analogue, which is desired in soft cheeses [137]. In another study on soy-based soft cheese analogues [28], the inclusion of G. candidum produced a softer, stickier texture compared to pure LAB starters (L. bulgaricus, St. thermophilus), which was attributed to the high degree of protein and fat degradation observed in the former samples. In this respect, cultures with strong enzymatic activities, especially proteolytic and lipolytic pathways, could be evaluated for the textural improvement of plant-based cheese analogues. EPS-producing LAB are potential candidates as well, and some strains have already been shown to improve structure in fat-reduced dairy cheeses [166,167].
Reproducing the authentic flavour of dairy-based cheeses is its own unique challenge. Proteins and fats in plant-based ingredients differ significantly from those found in animal milk [9,10,18]; naturally, most are perceived as off-flavours in the context of dairy cheeses. Ben-Harb et al. reported a reduction in green aldehydes during pea gel fermentation with cheese-isolated cultures [159]. They also observed that the overall volatile profiles differed substantially among gels made from pea PBAE, cow's milk, or a mixture of both, even when identical cultures were used, emphasizing the fact that cultures may display different fermentation characteristics in different matrices. Beyond this, none of the studies within the scope of our review specifically evaluated the reduction in off-flavours and/or the evolution of cheese and dairy-like notes in plant-based cheese analogues. Instead, sensory evaluation was performed to assess product likeability or to rate textural attributes; dairy benchmarks were rarely included [28,32,81].
Many fermented, plant-based cheese analogues on the market are made from cashews and almonds (Table 3). Unlike yoghurt analogues, a wide array of starter cultures is used in commercial cheese analogues, ranging from lactobacilli (direct inoculation or from yoghurt base) to water kefir and koji. Textural aids, however, are equally common in cheese analogue formulations. Interestingly, cauliflower and hemp serve as primary ingredients in a cheese sauce analogue developed by Grounded, which appears to use shio koji as its starter culture (Table 3). This deviates significantly from the overwhelming number of nut/drupe-based or occasionally grain-based cheese analogues on the market. With the exploration of non-dairy cultures and fermentation techniques, the market may soon welcome cheese and dairy analogues made from a greater variety of plant-based raw materials.

Kefir
Originally consumed in the region of Caucasia, kefir has exploded in popularity in recent years thanks to its various health benefits. Dairy kefir is traditionally made by fermenting milk with milk kefir grains, which owes its cauliflower-like appearance to the EPS, kefiran. A symbiotic consortium of LAB, yeasts, and acetic acid bacteria (AAB) is embedded in kefiran, which confers a sour, fizzy, and mildly alcoholic taste along with a viscous texture to fermented kefir. L. kefiranofaciens, L. kefiri, Lactococcus spp., Acetobacter pasteurianus, and Saccharomyces spp. have been identified as the major microbiota in milk kefir grains [136]. Water kefir grains, on the other hand, are used to ferment dairy-free, fruit-based sugar solutions to create a sparkling, acidic beverage. Instead of kefiran, their EPS is composed of α-glucans, and the fermentation conditions favour the predominance of a greater variety of AAB, yeasts (Saccharomyces spp., Dekkera bruxellensis), and the LAB species Lactobacillus nagelii and Lactobacillus hilgardii [168].
Many kefir fermentation studies with both milk and water kefir grains are directed at reaping its health benefits in a plant-based matrix instead of producing something that tastes and looks like authentic dairy kefir [169]. While there have been novel studies on the production of kefir-like products from walnut and soy PBAEs [170,171], flaxseed oil cake [172], and even apple juice [173], their organoleptic properties have not been evaluated against a dairy kefir benchmark. Hence, they may not possess the dairy organoleptic qualities required of a dairy analogue (i.e., the incubation of milk kefir grains in plant matrices does not necessarily result in a dairy kefir analogue).
A recent study by Yépez et al. evaluated milk kefir and water kefir grains in the fermentation of gelatinised flour suspensions made from oat, maize, or barley against a cow's milk control [124]. Though the study's main purpose was to determine the feasibility of in situ riboflavin fortification via co-fermentation of LAB with kefir grains, the viscosities and volatile profiles of the grain-based kefir analogues were also analysed. Milk kefir grains were found to present a better aptitude for lactic acid production and viscosity improvement, both quality indicators of dairy kefir. In particular, L. plantarum M5MA1-B2 further improved acetic acid and lactic acid content in maize kefir and oat kefir, respectively, and it enhanced viscosity when co-inoculated with water kefir, which highlighted the potential of using non-milk kefir strains in improving the quality of plant-based dairy kefir analogues.
There is a comparatively smaller offering of plant-based dairy kefir analogues on the market, most of which are made from coconut PBAE and require stabilisers in the form of starch. (Table 3). These include recent product launches, which are aligned with the new and growing interest in such analogues. Commercial dairy kefir analogues may, however, face stiffer competition compared to other dairy product analogues. As well as contending with their dairy counterparts, dairy kefir analogues also face existing competition in the plant-based sector in the form of water kefirs. Thus, organoleptic improvement is even more critical to engage and sustain consumer interest with a nutritious, tasty, and visually appealing product.

Targeted Fermentation and Precision Fermentation
Instead of the traditional fermentation of raw materials to create a plant-based dairy analogue, some researchers have used highly targeted fermentation approaches to eliminate or produce specific flavour compounds in plant-based raw materials, which may be further processed into dairy analogues or extracted to obtain flavour compounds of interest. For example, fungal fermentation of soy PBAE with Agrocybe aegerita successfully produced short chain fatty acids (SCFAs) reminiscent of Parmesan and Emmental cheese [174], while LAB and/or yeast fermentation of pea protein isolate was shown to reduce its green, leguminous, and bitter taste attributes, which are dairy-incompatible flavours [175][176][177]. In particular, Garcia Arteaga et al. reported the evolution of cheesy and salty notes after fermentation, though protein functionality was negatively affected [177]. Diacetyl and acetaldehyde, both important compounds in dairy-like aroma, are frequently reported in LAB fermentation of cereal-and soy-based matrices [33,65,178].
Precision fermentation processes have also been employed with genetically modified microorganisms (bacteria, yeast, or fungi) to synthesise dairy proteins and fats, which are then used as a base to create conventional dairy products. The reader is invited to refer to a recent review by Mendly-Zambo et al. [3] on this topic, which covers start-ups such as Perfect Day and the Real Vegan Cheese project [3]. Similar disruptors in this sphere include Change Foods, New Culture, Legendairy Foods, Better Dairy, and Remilk. The use of genetic modification (GM) technology, however, has always faced significant consumer resistance. Coupled with the relatively high cost of microbial engineering, much investment and research are expected before GM-produced dairy becomes commercially competitive against conventional dairy products or plant-based dairy analogues [3].

Conclusions
Plant-based dairy analogues represent a rapidly growing market segment, and the food industry has responded to consumer demand by developing a wide range of such dairy analogues from yoghurt to cheese, covering a range of plant materials from almonds to potatoes. The success of these products will greatly hinge on their organoleptic qualities in terms of aroma, taste, and texture, as well as other features such as stability and nutritional properties. The relative novelty of these products has resulted in the application of fermentation and other innovative processing methods on top of the exploration of a wide range of raw materials, consequently illustrating the vast options for scientists, technologists, and the industry. Fermentation has shown promising results in terms of imparting flavour and/or texture reminiscent of conventional dairy, especially in applications related to fermented dairy products such as yoghurt and cheese. The use of microorganisms not typically associated with dairy environments, e.g., EPS-producers, is also an avenue worth further exploration. Nonetheless, care needs to be taken to select and design products and processes that can match or surpass the organoleptic qualities of conventional bovine milk products so that they remain competitive in the market. Consequently, the application of technologies different from those used in the traditional dairy industry are expected since plant materials possess different physicochemical qualities and greater heterogeneity. It is anticipated that plant-based dairy analogues will continue to surge in popularity, and while organoleptic properties will remain the top factor in increasing consumer acceptability, greater efforts are warranted in making such products nutritionally whole and sustainable to produce, both for the wellbeing of consumers and the environment.
Author Contributions: A.P.: conceptualisation, methodology, investigation, writing-original draft, writing-review and editing, data curation, visualization. V.C.Y.T.: conceptualisation, methodology, investigation, writing-original draft, writing-review and editing, data curation. R.M.V.G.: conceptualisation, methodology, investigation, writing-original draft, writing-review and editing, data curation. J.S.: writing-review and editing, supervision, project administration. B.L.: conceptualisation, writing-review and editing, supervision, project administration, resources. S.Q.L.: writing-review and editing, supervision, project administration, funding acquisition. All authors have read and agreed to the published version of the manuscript.

Conflicts of Interest:
The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.