Amy Logan

Plant‐based milk alternatives (PBMAs) are now increasingly used by some consumers in similar ways to cow milk. However, they differ in their physicochemical and sensory properties. Improving PBMAs to better mimic dairy is limited by the time and cost of development trials. Machine learning offers a promising approach to predict optimal formulations without the need for extensive trial and error. In this study, Random Forest regression (RFR) and gradient boosting regression (GBR) models were used to formulate PBMAs (Formulation 1 and 2, respectively) that mimic cow milk. Sensory analysis, viscosity, stability, color, particle size, and headspace volatiles were compared to cow milk to assess model accuracy. Both formulations were sensorially distinct from coconut, oat, soy, almond, pasteurized, and ultra‐high temperature (UHT) cow milk, associated with vanilla, caramel, astringent, bitter, metallic, and nutty notes, compared to the sweet, dairy, and coconut notes of dairy. The formulations contained aldehydes, including hexanal and pentanal from soy, although compounds including vanillin, octanal, and 1‐pentanol were shared with cow milk. UHT cow (0.68 µm) had the smallest average particle size ( D ₅₀ ), followed by Formulation 2 and pasteurized cow (1.04 µm), with Formulation 1 (1.62 µm) as the largest. The UHT and pasteurized cow milk had higher whiteness indices compared with Formulation 1 and Formulation 2. Formulations exhibited faster sedimentation compared to cow milk. The models correctly predicted viscosity and some volatiles. However, predictive performance was limited by the small dataset and restricted access to commercial formulations. A larger dataset and access to formulations would improve model accuracy. Practical Applications: This research demonstrates the potential of predictive modelling for the development of plant‐based alternatives to animal‐derived ingredients. Such an approach would reduce reliance on trial and error in product development, saving time and development costs while meeting consumer needs.

Plant-based milk alternatives (PBMA) have emerged as popular substitutes for cow milk, driven by health, environmental, and ethical considerations. However, their ability to replicate the sensory and physicochemical properties of dairy remains a critical challenge for industry. This review critically examines the extent to which almond, soy, and oat PBMA replicate key sensory attributes of ultra-high temperature (UHT) full cream cow milk, focusing on appearance, texture, and flavour. Furthermore, it explores the relationship between these sensory attributes and the physicochemical properties of PBMA to elucidate the underlying reasons for the observed differences. A comparative analysis of compositional differences reveals fundamental limitations linked to plant protein functionality, carbohydrate structure, fat composition, and mineral fortification, all of which contribute to disparities in creaminess, mouthfeel, colour, and flavour. Technological strategies such as particle size reduction, enzymatic hydrolysis, and flavour masking have improved specific attributes, yet no PBMA fully replicates the holistic sensory experience of dairy. Emerging approaches, including blended formulations, precision fermentation, and artificial intelligence (AI)-driven optimisation, show promise in narrowing these gaps. Nonetheless, a complete replication of UHT cow milk remains elusive, highlighting the need for continued research and innovation to either approximate dairy properties more closely or enhance PBMA’s unique qualities to drive consumer acceptance.

The effects of κ‐casein (κ‐CN) and β‐casein (β‐CN) genetic variant and κ‐CN glycosylation degree (GD, low or high) on interfacial and foaming properties of bovine skim milk were investigated. No significant effect was measured for milks with different ĸ‐CN and β‐CN genetic variants. However, milks of higher GD exhibited lower surface tension, enhanced foamability and differences in secondary protein structure compared to lower GD skim milks. Glycan attachment is believed to affect surface activity and the spread and packing of protein at the foam bubble liquid–air interface, leading to differences in foaming performance.

Descriptive sensory analysis discriminated between the appearance, aroma, flavor, taste and texture of Cheddar cheese samples prepared from milk varying in fat globule size; small milk fat globule milk (SMM, d_4,3 = 2.76 ± 0.07 μm), large milk fat globule milk (LMM, d_4,3 = 5.07 ± 0.06 μm) and Control (d_4,3 = 3.91 ± 0.09 μm), over a six‐month maturation period. Significant differences were measured between SMM, LMM and the Control as each cheese matured, where SMM was the most different across all parameters examined. SMM was lowest in color intensity (34.6 < 38.8–47.0), firmness (29.4 < 36.1–47.7), and fracturability (24.5 < 28.8–30.7), and highest in cohesiveness (57.0 > 39.8–45.8), adhesiveness (52.9 > 39.0–45.4), and smoothness (50.4 > 37.9–44.0). Moreover, SMM was more intense than the Control in cultured odor (34.9 > 29.6–30.6) and flavor (45.5 > 34.9–36.2), rancid (23.3 > 16.0–18.6) and metallic (17.7 > 14.3) flavor, salty taste (45.7 > 40.3–40.5) and aftertaste (38.5 > 34.9–35.8), sour (36.3 > 28.9) and umami taste (19.6 > 16.4); attributes that align with the degree of maturation. LMM showed similar differences to the Control, but was firmer (36.1 > 29.4), less cohesive (45.8 < 57.0), adhesive (45.4 < 52.9) and smooth (44.0 < 50.4) and had a less intense cultured odor (31.5 < 34.9) than SMM. An increase in salty taste was associated with a decrease in cheese firmness (−0.91, p < 0.0001). These results could provide process‐driven solutions that allow the dairy industry to decrease maturation time or create differentiated cheese products to meet specific consumer tastes or function.