Orange Peel Feed Ingredient in Lactating Ewes: Effect on Yoghurt Chemical Composition, Fatty Acid Profile, Antioxidant Activity, Physicochemical Properties, and Sensory Quality

Featured Application

Incorporating orange peel into the diets of lactating ewes offers a sustainable and innovative solution for valorisation of by-products, making it a valuable approach for dairy production while addressing waste management challenges.

Abstract

This study examined the effects of incorporating processed (hydrolysed) (POP) and unprocessed (non-hydrolysed) (UOP) orange peels into the diets of lactating ewes on the proximate composition, fatty acid profile, antioxidant properties, physicochemical characteristics, and sensory attributes of traditional sheep milk yoghurt. Thirty-six Chios breed ewes were divided into three dietary groups: POP, UOP, and a conventional control diet (Control). Yoghurt produced from the UOP and POP diets had higher protein content (5.93 and 5.53%, respectively) and fat content (6.79 and 6.24%, respectively) compared to the Control (5.42% for protein and 6.06% for fat). Additionally, there were no significant differences in the nutritional indices of yoghurt fat. The determination of antioxidant activity showed no significant differences among the groups using the ABTS, DPPH, and FRAP methods, whereas significant differences were observed when measured with the Folin–Ciocalteu method. Textural analysis revealed that UOP yoghurt had greater firmness and cohesiveness. The whiteness index of yoghurt across all treatments remained consistent during storage, maintaining the desirable visual characteristics of traditional sheep milk yoghurt. Syneresis levels were similar across all groups, indicating no negative effects on yoghurt stability. Sensory evaluation confirmed the high acceptability of yoghurt produced from both experimental diets. These findings highlight the potential of orange peels as a sustainable feed additive, with no adverse effects on the nutritional, functional, or sensory properties of dairy products, while contributing to effective waste valorisation.

1. Introduction

The practice of feeding livestock with by-products from food processing industries dates back to the earliest days of animal domestication. This approach provides two key benefits: reducing livestock dependency on grains that could be used for human consumption and addressing the growing need to reduce the carbon footprint and increase sustainability in the food industry. The latter has become increasingly important due to population growth and the rising volume of crop and food by-products, especially in developed nations [1,2,3]. Additionally, the valorisation of agro-industrial by-products is necessary for global sustainability, as their traditional disposal in landfills presents significant environmental, social, and economic challenges. By-products such as peels, seeds, and husks hold high nutritional and functional value, making them valuable resources instead of waste. Biotechnological processes can transform these materials into new food ingredients, contributing to a circular bioeconomy that reduces waste, creates economic opportunities, and enhances human nutrition. Reinforcing their integration into the food supply chain promotes sustainable resource management while minimizing environmental impact [4,5]. Finally, the use of agricultural by-products, offers a potential way to lower feeding costs while maintaining feed efficiency [6,7].

Citrus fruit by-products, including sweet oranges, tangerines, lemons, and grapefruits, are well-suited for inclusion in ruminant diets due to the ability of ruminants to ferment high-fibre feed in their rumen. After juice extraction, the remaining residues consist of peel (flavedo and albedo), pulp (juice sac residue), rag (membranes and cores), and seeds [8,9]. Sweet oranges (Citrus sinensis Osbeck) constitute one of the world’s most popular fruit crops that, nowadays, are primarily used fresh or prepared as frozen juice concentrate [10].

In this context, orange peels represent a valuable by-product with significant potential for use in animal feed. As the primary by-product of the orange juice industry, they hold great promise as a secondary feed source. An orange consists of about 43% juice and 57% peel and pulp. For every litre of single-strength juice, 1.33 kg of peels are generated, while concentrated juice production yields more than 2.85 kg of peels per litre. Globally, orange peel production exceeds 8 million tonnes annually, with Europe contributing nearly 600,000 tonnes [3].

Various methods have been employed to improve the nutritional value of citrus by-products for animal feed. For example, fermenting citrus pulp, with or without molasses, has been effective in increasing its nitrogen content and buffering capacity. Similarly, ammoniation treatments, such as using urea or ammonium hydroxide, have significantly boosted the nitrogen content of dried citrus pulp. Other studies have explored supplementing non-protein nitrogen sources, such as urea or faba beans, to improve nitrogen utilisation and digestion. Among these, diets supplemented with faba beans demonstrated superior feeding value and animal performance compared to those supplemented with non-protein nitrogen alone. These techniques collectively enhance the nutrient profile of citrus by-products, making them a valuable resource for livestock nutrition [1].

In relation to orange peels, new methods are being developed to produce an improved feed ingredient by reducing pectic and free sugar content while increasing protein levels, resulting in a more balanced nutrient profile, as detailed in the recent study by Andrianou et al. [3].

The effect of feed composition on milk yield and composition has been widely studied [11,12,13]. Alternative feed can significantly impact the technological and coagulation properties of sheep and goat milk, which are influenced by milk fat and protein composition, as well as somatic cell count. Furthermore, incorporating alternative feeds into the animals’ diets can introduce nutraceutical or flavour-enhancing compounds into milk and dairy products [14]. The use of agro-industrial by-products and unconventional forages, rich in potentially anti-nutritional factors, can affect the quality of the raw milk and the dairy products made from it. Therefore, their inclusion should be carefully assessed to prevent any adverse effects on the quality and marketability of products like cheese and yoghurt. In this respect, the application of food industry by-products, such as sugar beet pulp, molasses, or tomato pomace, has affected the sensory characteristics, including the colour, firmness, and texture, as well as the fatty acid composition. However, these effects were neither consistent nor significant [15,16].

Regarding citrus by-products in sheep rations, no adverse effects on milk production or composition have been reported. For example, Fegeros et al. [17] used dried citrus pulp to partially replace concentrate feed and found no significant effect on milk yield or composition. However, it did reduce the saturated fatty acids (C4:0–C10:0). In that study, the entire orange pulp, including seeds and peels, was dried after juice extraction without undergoing depectinisation or molasses extraction. The first step in processing dried citrus pulp involves adding lime to reduce its hydrophilic properties, which are attributed to the presence of pectin. Additionally, Volanis et al. [18] found that ensiled sliced oranges made from non-marketable oranges can be fed to lactating ewes without negative effects on their performance. However, the impact of such diets on the composition, physicochemical, and sensory characteristics of dairy products (cheese and yoghurt) has not been sufficiently documented.

Produced through the fermentation of Lactobacillus delbrueckii subsp. bulgaricus and Streptococcus thermophilus, yoghurt is one of the most widely recognised and consumed cultured milk products globally. It has been a significant part of the human diet for thousands of years. In Greece, sheep milk yoghurt is particularly popular due to its firm, smooth texture, distinctive sensory qualities, and the abundant supply of milk from the country’s large dairy sheep population [19,20].

This study aims to evaluate the impact of incorporating orange peel as a feed ingredient on the quality of sheep milk yoghurt, contributing to sustainable dairy production.

2. Materials and Methods

2.1. Diets and Animals

Three isonitrogenous and isoenergetic diets were formulated to meet the nutrient requirements of Chios breed ewes during lactation [21]. The Control (conventional) diet consisted of corn grain, barley grain, wheat grain, soybean meal, sunflower meal, and a vitamin and mineral premix. In the experimental diets, wheat grain and sunflower meal were partially replaced with dried orange peels, as detailed in

Table 1. Ingredient composition of the experimental diets.

Ingredient (g/kg)	Diet
	Control	UOP	POP
Corn grain	300	300	300
Barley grain	200	200	200
Wheat grain	200	120	120
Soyabean meal	110	110	110
Sunflower meal	150	120	120
Orange peels	0	110	110
Limescale	5	5	5
Monocalcium phosphate	5	5	5
Salt	5	5	5
Vitamin and mineral premix	25	25	25

UOP = Unprocessed orange peels; POP = Processed orange peels.

.

The production process of the orange peels has been described in detail by Andrianou et al. [3]. Briefly, orange peels were hydrolysed in a pilot bioreactor, after which the released sugars were aerobically fermented to produce yeast. The hydrolysed peels and produced yeast were stabilised in a pilot drying plant. Both processed (hydrolysed) and unprocessed (non-hydrolysed) orange peels were dehydrated and milled simultaneously into a homogeneous coarse powder using a demo-scale rotary drum dryer with a biomass burner. This process, performed at 120 °C for 15–20 h per cycle, aimed to avoid microbial development while preserving feedstock composition. Τhe optimised conditions for the saccharification (Saccharomyces cerevisiae yeast) process were 50 °C, 24 h, 7.5% solids loading, Pectinex 25 μL/g TS, CellicCTec3 25 μL/g TS.

The enzymatic hydrolysis of orange peels aimed to produce an advanced animal feed ingredient with higher protein content. The process yielded two fractions: a liquid fraction rich in sugars and a hydrolysed solid residue. The liquid fraction was further utilised for yeast cultivation, producing single-cell protein [3].

A total of 36 healthy Chios breed ewes in their 2nd and 3rd lactation were randomly assigned to one of the three dietary groups, with 12 ewes per group. These groups were fed the formulated diets during the study. Sheep were fed the experimental and control diets for 28 days before milk was collected for the first batch of yoghurt production.

Concerning the use of 36 dairy sheep, the minimum number of animals to be sampled was calculated—sufficient to ensure the statistical significance and the validity of the results obtained, but not more than necessary which is fully consistent with the 3Rs concept (“reduction, refinement, and replacement”)—obtaining new knowledge without the need for the use of additional experimental animals. Prior to the onset of the experiment, the minimum required total sample size was calculated using the General Linear Multivariate Model with Wilks Likelihood Ratio procedure in the GLIMMPSE platform “https://glimmpse.samplesizeshop.org/, RRID: SCR_016297, (accessed on 20 December 2022)” [22]. The desired power was set at 0.8, the type I error rate at 0.05 and the results of the analysis revealed that a minimum sample size of 36 sheep (12 per group) was required.

2.2. Milk Collection and Yoghurt Production

Animals were milked twice a day. For yoghurt production, milk from each group (evening and morning milk) was collected separately in designated tanks in the milking parlour and then combined into a single container per group. Milk from the evening and morning milking within the same group was thoroughly mixed. Shortly after the morning milking, approximately 10 L of milk from each group was transported under refrigerated conditions on the same day.

The distance between Paralimni Giannitsa (milking parlour) (40°45′5.0904″ N 22°27′10.9548″ E) and Florina (laboratory) (40°46′56.2764″ N 21°24′35.388″ E) is approximately 120 km, with an average travel time of 90 min. Upon arrival at the laboratory, the milk was poured through an aluminium milk strainer equipped with a filter paper (disk filters for cans, ∅170 mm, Albert Kerbl GmbH, Buchbach, Germany) suitable for milk aimed for human consumption to remove large debris, such as bedding, hair, and soil parts, and placed in the refrigerator. The milk was also examined for the presence of antibiotic residue to ensure its suitability for yoghurt production. The EcoTest Easy MRL^® (Acquanegra Cremonese, Italy) was employed for this analysis. The examination for the presence of antibiotic residue was precautionary, since the milk was obtained from the flock raised at HAO-Demeter, where information on all veterinary interventions was known.

The milk was pasteurised (85–90 °C for 30 min), then cooled to a temperature range of 40–45 °C in an ice bath. When it reached approximately 42 °C, the starter culture was prepared [23]. During milk cooling, the effectiveness of pasteurisation was evaluated by the rapid alkaline phosphatase test using the Lactognost test kit (Heyl Chemisch-pharmazeutische Fabrik GmbH & Co. KG, Berlin, Germany) [24]. For the first batch of yoghurt production, a commercially available traditional yoghurt made from sheep’s milk was used as the starter culture. For the subsequent three batches, a yoghurt sample from the previous batch served as the starter culture. The starter culture was prepared by diluting sheep milk yoghurt into the milk at a concentration of 3% w/v, similar with the culture percentage used in the study by Serafeimidou et al. [25]. To ensure a more uniform dilution, the starter culture was first mixed with a small portion of pasteurised milk (approximately 200 mL). This mixture was then incorporated into the remaining milk through gentle stirring. The resulting mixture was transferred into designated containers based on the type of analysis. For proximate composition, titratable acidity, pH, and syneresis by gravity, 125 mL transparent PET (polyethylene terephthalate) containers with lids were used. Sensory evaluation samples were prepared in similar but smaller size (70 mL) containers. Fatty acid composition, total phenolic content, and antioxidant profile samples were placed in 50 mL plastic centrifuge tubes with lids. Samples destined for viscosity determination were placed in 300 mL transparent PP (polypropylene) containers (beakers) with lids. Finally, colour and texture analysis samples were placed in custom-made white containers (height 60 mm and diameter 50 mm). These containers were fabricated using a 1.75 mm filament (Neema3D, Athens, Greece) with a CreatBot D600 Pro 3D printer (Henan Creatbot Technology Limited, Zhengzhou, China). They were designed to replicate the containers specified for the back extrusion tests of semi-liquids during texture analysis. During refrigerated storage, and in order to prevent surface drying, the latter containers were covered with commercial kitchen cling (stretch) film and stored in a plastic container with a lid.

All containers were placed in an incubator (Incuterm Digit ID-150, RAYPA, R. Espinar S.L., Tarrasa, Spain), set to 45 °C until the pH reached 4.5 [23]. During fermentation, pH levels were monitored at regular intervals of 30–45 min. Once fermentation was complete, the yoghurt was left to cool at room temperature on the bench before being stored in a domestic refrigerator for 21 days. A shelf life of 22–25 days is considered to be the safe consumption period for traditional types of yoghurt [26,27].

The produced yoghurt was classified as traditional, since they meet the following specifications required by Greek food legislation: (a) it is produced using the traditional method so that it has a crust (membrane) on its surface; (b) it results from the coagulation of exclusively fresh or pasteurised milk which has not undergone any modification of its natural composition, with the sole exception of the adjustment of fat content, up to the point where it is technically possible to produce a crust [28]. The formation of crust is due to the fact that the milk has not been subjected to standardisation and homogenisation treatments regarding its fat content [25].

The production process for traditional yoghurt is illustrated in Figure 1. Four batches of yoghurt were prepared using milk collected on four different days throughout the lactation period, with the sampling conducted between March and May 2023.

The three types of yoghurt produced were labelled as Control, UOP, and POP, based on the type of feed consumed by the ewes whose milk was used.

2.3. Microbiological Analyses

Microbiological analyses were conducted on samples collected on storage day 2, with duplicate analyses performed for each of the three samples from each production batch. The following microorganism groups were examined: (i) total coliforms in Crystal Violet Neutral Red Bile Lactose (VRBL) Agar at 37 °C for 24 h [29]; (ii) Staphylococcus aureus on Baird Parker Agar at 37 °C for 48 ± 4 h, and confirmed by a positive coagulase test [30]; (iii) yeast and mould on the yeast extract/dextrose/chloramphenicol/agar medium at 25 °C for 5 days [31]; (iv) b-glucuronidase-positive Escherichia coli in a Tryptone bile x-glucuronic medium (TBX) at 44 °C for 18–24 h. All bacteria groups were enumerated using the pour plate method, except for Staphylococcus aureus, which was enumerated using the spread plate method.

Additionally, the detection of Salmonella spp. in 8 samples from each production batch was performed using the culture method specified in ISO 6579-1:2017 [32]. The procedure involved four successive stages: (i) pre-enrichment in Buffered Peptone Water at 35 °C for 18 h; (ii) selective enrichment in Rappaport–Vassiliadis Soya Broth (RVS broth) (Oxoid, Hampshire, UK, CM0866) at 41.5 °C for 24 h and in Müller–Kauffmann Tetrathionate–Novobiocin Broth (MKTTn broth) at 37 °C for 24 h; (iii) plating onto selective solid media, including Xylose Lysine Deoxycholate Agar (XLD agar) and Brilliant Green Agar, both incubated at 37 °C for 24 h; and (iv) subculturing presumptive Salmonella colonies, followed by confirmation of their identity through appropriate biochemical and serological tests.

Following the incubation period, colonies were enumerated, and the results were expressed as the logarithm of the mean colony-forming units i.e., log₁₀ CFU/g, except for Salmonella spp., which was reported solely as detected or not detected. Microbial analyses were performed prior to conducting the sensory evaluation of the products.

2.4. Determination of Physicochemical Characteristics (Milk) and Proximate Composition (Milk and Yoghurt)

All analyses of milk composition and physicochemical properties were performed on the day of milk collection. Milk composition (protein, fat, lactose, and total solids) and freezing point depression (FPD) was measured using MilkoScan FT 6000 equipment (Foss, Hillerød, Denmark). Milk physicochemical characteristics, such as pH, electrical conductivity, refractive index and Brix value, were measured as described in the study of Kasapidou et al. [33]. Prior to analysis, samples were placed in a controlled-temperature water bath to reach room temperature (20 °C) and were then gently mixed by inverting the container multiple times to ensure homogeneity without causing frothing. The milk pH was measured using a glass electrode with a built-in temperature sensor (5014T, Crison Instruments, s.a., Barcelona, Spain) in a pH-meter (GLP 21, Crison Instruments, s.a., Barcelona, Spain), calibrated with standard buffer solutions of pH 4.0 and 7.0 according to the manufacturer’s instructions. Milk titratable acidity was determined according to the AOAC 947.05 method [34]. The electrical conductivity of the samples was measured using a conductometer (GLP 31, Crison Instruments, s.a., Barcelona, Spain) and a Sodium Ion-Selective Electrode (50 70, Crison Instruments, s.a., Barcelona, Spain), calibrated with 147 μS/cm, 1413 μS/cm, and 12.88 mS/cm buffer solutions. Refractive Index and Brix value were determined using a digital refractometer equipped with a Peltier thermostat (DR6000-T, Krüss GmbH, Hamburg, Germany).

The proximate composition of the yoghurt samples was determined according to the recommended standard methods applied in milk samples [34]. The samples were collected on storage day 7 and analysed within one week following collection. Before analysis, samples were thoroughly but gently homogenised by hand with a spatula. This was particularly important, since the yoghurt was produced from non-homogenised milk that leads to the formation of a crust. Gentle mixing was required to prevent extensive whey separation during homogenisation.

Moisture and total solid content was determined by the 990.20 method after drying of the homogenised sample in a convection chamber (ED-115, Binder GmbH, Tuttlingen, Germany) at 102 °C until a constant weight was obtained. Dishes were placed in a desiccator and weighed when they obtained room temperature. Moisture was reported as weight loss from pre- and post-drying weights. Total solid content was calculated as a percentage of the post-drying weight.

Ash content was determined using the 935.42 method. Samples were incinerated at 550 °C for 12 h in a muffle furnace (model LM 412.07, Linn High Therm GmbH, Eschenfelden, Germany) until light grey ash was obtained. Ashing dishes were placed in a desiccator and weighed after reaching room temperature. Ash content was calculated by weight loss before and after incineration.

Protein content was determined according to the method 991.20 by using nitrogen digestion (Turbotherm type TT/12M) and distillation (Vapodest type 40) apparatuses (C. Gerhardt Fabrik und Lager chemischer Apparate GmbH & Co. KG, Königswinter, Germany) and converted to crude protein by multiplying the nitrogen content by 6.38.

Fat content was measured with a modification of the van Gulick method as follows [35]: a homogenised sample was weighed in the weighing beaker of a cheese butyrometer, and sulfuric acid was added until the sample was covered. The butyrometer was kept in a water bath (65 °C) following the addition of amyl alcohol and sulfuric acid. The butyrometer was centrifuged in a thermostatically controlled Gerber centrifuge (Micro II, Gerber Instruments AG, Effretikon, Sweden) and finally the fat content was read on the butyrometer scale.

The Solids-not-Fat (SnF) content was determined by subtracting the percentage of fat from the percentage of the total solids according to the 990.21 method [34].

The content of carbohydrate was calculated by subtracting the sum percentage of moisture, ash, protein, and fat from 100.

2.5. Determination of Fatty Acid Composition and Nutritional Indices

Yoghurt samples that were collected on storage day 7 were frozen and stored for analysis, which was conducted within six months following collection. The samples were thawed overnight at 4 °C, and the following day, fatty acids were extracted according to the method of O’Fallon et al. [36]. Briefly, samples were thoroughly mixed with a spatula before transferring a portion into a screw-capped Pyrex tube. Aqueous potassium hydroxide solution and methanol containing t-butyl-hydroxytoluene (BHT) were added to prevent oxidation of unsaturated fatty acids. The tubes were heated in a water bath and periodically shaken to ensure proper mixing and hydrolysis. After cooling, aqueous sulfuric acid was added, and the mixture was gently inverted before reheating and shaking again. Once cooled, hexane was added, and the tubes were vortexed to complete the process. Following that, the upper hexane phase was filtered through a 0.45 µm pore size PVDF syringe filter, transferred into amber GC vials, and stored at −20 °C until analysed.

Fatty acid methyl ester analysis was performed on an Agilent Technologies 6890N GC (Agilent Technologies, Inc., Santa Clara, CA, USA) equipped with a flame ionisation detector (FID) and a 60 m × 0.25 mm i.d., 0.25 μm film thickness DB-23 (50% cyanopropyl 50% dimethyl polysiloxane) capillary column (Model Number: Agilent 122 2362). The injector temperature was set at 250 °C. The oven temperature was programmed from 110 °C (held for 6 min), to 165 °C at 1 °C/min (held for 13 min), to 195 °C at 15 °C/min (held for 22 min) and to 230 °C at 7 °C/min (held for 7 min). The carrier gas was helium at 0.7 mL/min and the injection volume was set at 3 μL and the E Split ratio was 1:50. The injection was performed using an Agilent 7683 Series auto-sampler. Fatty acids were identified using the information of the retention times of the three commercial standard mixtures: (a) 37 component FAME mix (Supelco^®, 47885-U, Sigma-Aldrich, Taufkirchen, Germany), (b) PUFA-2, Animal source (Supelco^®, 47015-U, Sigma-Aldrich, Taufkirchen, Germany), and (c) a mixture of cis- and trans-9,11- and -10,12-octadecadienoic acid methyl esters (Sigma, O5632, Sigma-Aldrich, Taufkirchen, Germany). Fatty acids were quantified by peak area measurement and the results are expressed as percent (%) of the total peak areas for all quantified acids. Fatty acids were grouped as saturated fatty acids (SFA), monounsaturated fatty acids (MUFA), polyunsaturated fatty acids (PUFA), and unsaturated fatty acids (UFA).

The fatty acid profile was used to determine the nutritional indices associated with healthy fat intake. The applied indices are those outlined in the recent review by Chen and Liu [37] for ruminant dairy products. These researchers systematically compiled information on fatty acid profiles published since 2000 to enhance the understanding of the implications and applications of diverse nutritional indices.

$$\begin{gathered} \mathrm{Atherogenicity\; Index} \\ \mathrm{AI}={\displaystyle \frac{[\mathrm{C}12:0+(4\times \mathrm{C}14:0)+\mathrm{C}16:0]}{\Sigma \mathrm{U}\mathrm{F}\mathrm{A}}} \end{gathered}$$

$$\begin{gathered} \mathrm{Thrombogenicity\; Index} \\ \mathrm{TI}={\displaystyle \frac{(\mathrm{C}14:0+\mathrm{C}16:0+\mathrm{C}18:0)}{\left[\right(0.5\times \Sigma \mathrm{M}\mathrm{U}\mathrm{F}\mathrm{A})+(0.5\times \Sigma \mathrm{n}-6\mathrm{P}\mathrm{U}\mathrm{F}\mathrm{A})+(3\times \Sigma \mathrm{n}-3\mathrm{P}\mathrm{U}\mathrm{F}\mathrm{A})+(\mathrm{n}-3/\mathrm{n}-6\left)\right]}} \end{gathered}$$

$$\begin{gathered} \mathrm{Hypocholesterolaemic:\; hypercholesterolaemic\; fatty\; acid\; ratio\; (h/H)} \\ \mathrm{h}/\mathrm{H}={\displaystyle \frac{(\mathrm{C}18:1\mathrm{n}-9\mathrm{c}\mathrm{i}\mathrm{s}+\Sigma \mathrm{P}\mathrm{U}\mathrm{F}\mathrm{A})}{(\mathrm{C}12:0+\mathrm{C}14:0+\mathrm{C}16:0)}} \end{gathered}$$

$$\begin{gathered} \mathrm{Health-promoting\; index} \\ \mathrm{HPI}={\displaystyle \frac{\Sigma \mathrm{U}\mathrm{F}\mathrm{A}}{\left[\mathrm{C}12:0+4\times \mathrm{C}14:0\right)+\mathrm{C}16:0]}} \end{gathered}$$

$$\begin{gathered} \mathrm{Polyunsaturated\; fatty\; acid/Saturated\; fatty\; acid\; ratio} \\ \mathrm{PUFA}/\mathrm{SFA}={\displaystyle \frac{\Sigma \mathrm{P}\mathrm{U}\mathrm{F}\mathrm{A}}{\Sigma \mathrm{S}\mathrm{F}\mathrm{A}}} \end{gathered}$$

Finally, the desirable fatty acid (DFA) index was calculated according to the formula of Rhee [38] as follows:

DFA = ΣUFA + C18:0

2.6. Determination of Total Phenolic Content and Antioxidant Profile

Frozen yoghurt samples collected on day 7 of storage were analysed within six months. Before analysis, the samples were thawed overnight. A homogenised yoghurt sample (10 g) was then mixed with 3 mL of distilled water in a 50 mL screw-capped centrifuge tube and vortexed thoroughly to prepare an aqueous extract for determining the total phenolic content and antioxidant profile, following a modified method based on Perna et al. [39]. The mixture was centrifuged at 7000 rpm for 20 min at 20 °C using a refrigerated centrifuge (Sorvall RC6 Plus, Thermo Fisher Scientific, Langenselbold, Germany). The resulting supernatant was filtered through a 0.45 µm PVDF syringe filter and stored in 1.5 mL aliquots in Eppendorf^® tubes at −20 °C until analysis [40,41].

The antioxidant activity was assessed using the Folin–Ciocalteu, DPPH, FRAP, and ABTS methods, to evaluate various aspects of the antioxidant properties of the yoghurt. The Folin–Ciocalteu method determined the total phenolic content (TPC), expressed as mg of gallic acid equivalents (GAE) per g of yoghurt [42]. The DPPH (2,2-diphenyl-1-picrylhydrazyl) method assessed the free radical scavenging activity, following the protocol of Sanchez-Moreno et al. [43] with slight modifications. The results were expressed as μΜ of Trolox equivalents (TE) per g of yoghurt, where Trolox is a water-soluble analogue of vitamin E: (6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid). The FRAP (ferric reducing antioxidant power) method evaluated the reducing power activity of the samples, as reported by Pulido et al. [44], with minor changes. The results were expressed as μΜ of Trolox equivalents (TE) per g of yoghurt. Finally, the ABTS [2,2′-azinobis-(3-ethylbenzothiazoline-6-sulfonic acid)] method examined the total antioxidant capacity of the samples, following the procedure outlined by Re et al. [45], with slight adjustments. The results were expressed as μΜ of Trolox equivalents (TE) per g of yoghurt.

2.7. Determination of Yoghurt Active (pH) and Titratable Acidity

Yoghurt active (pH) acidity was determined on storage days 7, 14, and 21, while titratable acidity was measured on storage day 0. Yoghurt pH was determined on a 10 g homogenised sample that was dissolved in 100 mL of distilled water. The pH was measured using a glass electrode with a built-in temperature sensor (5014T electrode, Crison Instruments, s.a., Barcelona, Spain) in a Crison GLP 21 pH-meter (Barcelona, Spain) which was calibrated with standard buffer solutions of pH 4.0 and 7.0 according to the manufacturer’s instructions.

The titratable acidity was determined with a modification of the AOAC 947.05 method [34]. Briefly, approximately 10 g of the sample was dissolved in 100 mL distilled water and mixed thoroughly. A few drops of phenolphthalein indicator were added, and it was titrated against a standardised 0.1N sodium hydroxide solution until a pale pink colour persisted for about 10–15 s for complete neutralisation. The titratable acidity was expressed as g of lactic acid per 100 g of yoghurt, since lactic acid is the main organic acid in yoghurt samples.

2.8. Instrumental Colour Measurement

Colour measurements were conducted on storage days 7, 14, and 21 in three samples from each production batch. Colour measurements were carried out with the Minolta CR-410 colorimeter (Konica Minolta Company, Osaka, Japan) with a 50-mm measuring area (aperture size) using illuminant source C and 2° standard observer angle. The light projection tube (CR-A33a, Konica Minolta, Japan) was applied over the aperture port while measuring. Initially, before the measurement, calibration was performed using a white calibration plate (Y = 93.9, x = 0.3136, y = 0.3201). The measurements were conducted in triplicate on the surface colour of intact yoghurt samples. The measurements were recorded using the SpectraMagic™ NX software (Lite 3.4 version) (Konica Minolta Company, Osaka, Japan). Changes in colour and texture during storage were evaluated using the same yoghurt samples prepared in custom-made white containers (Figure 2).

The colour measurement was conducted using the CIE Lab* system. In addition to the L*, a*, and b* parameters, chroma (colour saturation), hue, and the whiteness index values were determined according to the following equations, as reported by Milovanovic et al. [46], for chroma and the whiteness index, and McLellan et al. [47], for the hue:

$$\mathrm{Chroma}=\sqrt{\left({\mathrm{a}}^{*2}+{\mathrm{b}}^{*2}\right)}$$

$$\mathrm{Hue}\mathrm{angle}={\mathrm{t}\mathrm{a}\mathrm{n}}^{-1}\left({\displaystyle \frac{\mathrm{b}\ast }{\mathrm{a}\ast }}\right)\times \left({\displaystyle \frac{180}{\mathsf{\pi }}}\right)$$

Whiteness index = 100 − ((100 − L*)² + a*² + b*²)^0.5

Total colour difference (ΔE_Lab) between the Control (C) and samples from the UOP and POP treatments within the same storage day was also calculated according to the following equation [46]:

$${\mathsf{\Delta }\mathrm{E}}_{\mathrm{Lab}}=\sqrt{({\mathrm{L}*}_{\mathrm{C}}-{\mathrm{L}*}_{\mathrm{T}}{)}^2+({\mathrm{a}*}_{\mathrm{C}}-{\mathrm{a}*}_{\mathrm{T}}{)}^2+({\mathrm{b}*}_{\mathrm{C}}-{\mathrm{b}*}_{\mathrm{T}}{)}^2}$$

where C = Control and T = treated samples i.e., UOP and POP.

2.9. Texture Analysis

The yoghurt consistency was examined using the back extrusion test, in which the sample was penetrated with a cylindrical plate probe. Analyses of firmness, adhesiveness, and stickiness were conducted on storage days 7, 14, and 21. The Perten TVT 6700 texture analyzer and TexCal5^® instrumental software (version 5.3.5.36) (Perten Instruments, Hägersten, Sweden) were employed according to the method 41.01 (Yoghurt consistency) with a double cycle analysis. Compression plate P-CP35S with a 35 mm diameter and a 10 kg load cell was employed in the procedure.

The viscosity of the samples was measured at 20 °C using Viscometer Visco Star plus (FUNGILAB, S. A., Barcelona, Spain) with stainless steel spindle R4 at a steering rate of 30 rpm according to manufacturer’s guidelines for spindle selection. An average of 10 measurements was taken and the results were expressed in mm²/s (kinetic viscosity based on product density). Density (g/mL) was determined by filling a pre-weighed 60 mL medical syringe cylinder, suitable for parenteral feeding, with yoghurt and recording the weight.

The temperature of the yoghurt samples was maintained consistently during both tests. Specifically, the samples were allowed to equilibrate at room temperature for two hours before the measurements.

2.10. Determination of Yoghurt Syneresis by Gravity and by Centrifugation

Syneresis of the yoghurt was determined using both the gravity (drainage) method and the centrifugal method with modifications of the methods presented by Harwalkar and Kalab [48] on storage days 7, 14, and 21. Regarding the drainage method, a cylindrical sample with a volume of approximately 5 mL (obtained with a special spoon) of the yoghurt was placed with its flat side onto a sieve (mesh width of 1.18 mm) and the drained whey was collected in pre-weighed centrifuge (15 mL) test tubes, as shown in Figure 3. The weight of the yoghurt sample was also recorded. The set consisting of the sieve, the spoons with the yoghurt samples, the funnels, and the centrifuge tubes was placed in a RAYPA IRE-160 refrigerator incubator (R. Espinar S.L., Tarrasa, Spain) set at 10 °C and the amount of drained off whey was measured after 2 h.

When syneresis was determined using the centrifugal method, a pre-weighed centrifuge tube (50 mL) was used as the container for the yoghurt samples during preparation, incubation, and storage. On the examination day, the tube was centrifuged at 4000 rpm for 10 min at a temperature of 10 °C in a refrigerated centrifuge (Sorvall RC6 Plus, Thermo Fisher Scientific, Langenselbold, Germany). The supernatant was poured into a Pyrex beaker and the weight was recorded. Yoghurt syneresis was expressed as a percentage of the initial yoghurt weight in both methods.

2.11. Sensory Analysis

A student panel consisting of eight members assessed the yoghurt samples. Sensory analysis was conducted on samples stored for seven days. All students reported frequent consumption of sheep yoghurt. Additionally, they participated in a training session designed to familiarise them with the evaluation procedure and questionnaire. During the session, all questions were clarified, and test trials were conducted using commercial yoghurt. The samples were coded with a three-digit random number (Figure 4). The presentation order of samples was randomised across participants. The evaluation of the products included overall appearance, colour, aroma intensity, aroma, taste intensity, taste, acidity, texture (spoon), texture (mouth), syneresis, aftertaste, and overall acceptability using a 7-point hedonic scale, as presented in

Table 2. Sensory evaluation descriptors of the yoghurt samples.

Characteristic	Hedonic Scale Descriptor
	1	4	7
Overall appearance
(colour, surface smoothness i.e., presence of lumps, whey separation and presence of air pockets)	Unacceptable	Acceptable	Exceptional
Colour	Unacceptable	Acceptable	Exceptional
Odour intensity	Extremely weak	Average	Extremely intense
Odour	Unacceptable	Acceptable	Exceptional
Taste intensity	Extremely weak	Average	Extremely intense
Taste	Unacceptable	Acceptable	Exceptional
Acidity	Extremely weak	Average	Extremely intense
Texture (spoon)	Unacceptable	Acceptable	Exceptional
Texture (mouth)	Unacceptable	Acceptable	Exceptional
Syneresis (at spoon cutting)	Minimum	Average	Intense
Aftertaste	Minimum	Average	Intense
Overall acceptability	Unacceptable	Acceptable	Exceptional

. Yoghurt samples were removed from the refrigerator 30 min before the test and left at room temperature.

2.12. Statistical Analysis

The data were presented as treatment means ± standard error of the mean (SEM). Variance homogeneity was evaluated using the Levene test. When variances were homogeneous, one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test was conducted. For cases where homogeneity of variances was not achieved, the Games–Howell test was applied to identify differences between the treatments. Changes within the same treatment during storage were also evaluated for colour, texture, viscosity, and syneresis. Statistical significance was established at a threshold of p < 0.05 for all analyses. Data processing was carried out using SPSS software (version 29.0, SPSS Inc., Chicago, IL, USA).

3. Results and Discussion

3.1. Milk Preparation for Yoghurt Production

Raw milk was tested for antibiotic residue, and no residue was detected, ensuring the milk met the necessary quality standards for yoghurt production. Pasteurisation was then carried out, and its effectiveness was confirmed with a rapid alkaline phosphatase test. These steps guaranteed the production of safe and high-quality yoghurt.

The presence of drug residue in milk can pose risks, such as allergies and microbial resistance, and may negatively affect the fermentation process by inhibiting starter culture activity [49,50]. In this study, the presence of antibiotics in sheep milk yoghurt led to significant delays in coagulation time, impacting the physicochemical and organoleptic properties, and ultimately lowering commercial quality and causing potential economic losses [49].

Heat treatment of the milk serves to eliminate interfering microorganisms, to denature whey proteins, and to enhance the texture of the final product. It also improves the milk’s appearance and increases viscosity, while causing flavour changes and a slight reduction in nutritional value [51,52]. The heat treatment conditions applied (80 to 85 °C for 30 min, or 90 to 95 °C for 5 min) are considered adequate for producing high-quality yoghurt [53]. Proper heat treatment is crucial for achieving the desired yoghurt texture, ensuring better gel strength and firmness, and reducing the risk of whey separation [54].

3.2. Yoghurt Microbiological Evaluation

The microbial quality of the yoghurt was assessed prior to sensory evaluation by the taste panellists. As shown in

Table 3. Yoghurt microbial quality.

Microorganism	Treatment
	Control (n = 12)	UOP (n = 12)	POP (n = 12)
Coliform spp. (log10 CFU/g)	<DL	<DL	<DL
Staphylococcus aureus (log10 CFU/g)	<DL	<DL	<DL
Escherichia coli (log10 CFU/g)	<DL	<DL	<DL
Moulds and yeasts (log10 CFU/g)	<DL	<DL	<DL
Salmonella spp. (Presence/Absence)	Absence	Absence	Absence

UOP = Unprocessed orange peels; POP = Processed orange peels; DL = detection limit; DL = 1.0 log CFU/g for Coliform spp., Escherichia coli, moulds and yeasts (pouring inoculation method) and 2.0 log CFU/g for Staphylococcus aureus (spreading inoculation method).

, no microorganisms were detected in any of the yoghurt samples from all treatments and production batches, with no colony-forming units (CFU/g) observed. Additionally, all samples tested negative for Salmonella spp. The results regarding the microbial quality of the yoghurt are related to the effective milk pasteurisation as well as the preparation of yoghurt under strict hygienic conditions that prevented contamination during and after production [55,56]. Based on the results of Coliform spp., moulds, and yeasts, yoghurt quality is considered to be satisfactory. Additionally, yoghurt acidity that was higher than 1g lactic acid/100 g did not allow the growth of Salmonella spp. [57]. Furthermore, Staphylococcus spp. are usually unlikely to survive during fermentation [57], whereas acid adaptation of Escherichia coli in fermented milk products as well as reported outbreaks of Escherichia coli O157 with the consumption of yoghurt prompt for the examination of dairy products [57,58].

3.3. Milk and Yoghurt Proximate Composition

Milk composition as well as physicochemical characteristics are presented in

Table 4. Milk proximate composition and physicochemical characteristics.

Component	Treatment			SEM	Significance
	Control (n = 4)	UOP (n = 4)	POP (n = 4)
Protein (%)	5.27	5.58	5.31	0.083	NS
Fat (%)	5.91	6.77	6.25	0.169	NS
Lactose (%)	4.66	4.72	4.78	0.024	NS
Total solids (%)	16.68 a	17.90 b	17.17 ab	0.186	**
FPD (-°C)	0.542 b	0.561 a	0.557 a	0.027	**
pH	6.73	6.78	6.71	0.014	NS
Titratable acidity (g lactic acid/100 g milk)	0.148	0.146	0.153	0.004	NS
Electrical Conductivity (mS/cm)	3.92	4.03	3.99	0.022	NS
Refractive index (nD20)	1.3531	1.3539	1.3537	0.000	NS
Brix (°Bx)	13.38	13.88	13.93	0.104	NS

UOP = unprocessed orange peel; POP = processed orange peel; ** = p < 0.01; NS = non-significant; superscripts a, b differ at p < 0.05.

. There were no significant differences (p > 0.05) in the contents of protein, fat and lactose whereas highly significant differences (p < 0.01) were observed in the total solids content. These differences are related to the content of ash indicating that inclusion of either unprocessed or processed orange peels (UOP and POP treatments) led a higher ash content.

The protein and fat content of the milk were similar to the values reported by the Hellenic Agricultural Organisation DIMITRA [59] for raw sheep milk produced both in the Regional Unit of Pella and across the entire country during the study period (March–May 2023). Specifically, protein content was 4.36 g/100 g in the Regional Unit of Pella and 4.35 g/100 g for the entire country, while fat content was 5.28 g/100 g and 5.33 g/100 g, respectively. In terms of lactose content, all treatments showed higher levels compared to the milk produced in the Regional Unit of Pella (4.59 g/100 g). Notably, milk from the UOP and POP treatments exceeded the lactose content reported for milk produced nationwide (4.67 g/100 g). The milk composition fell within the range reported by Pappa et al. [60] for milk analysed both at the farm and at the dairy plant in northwestern Greece. It was also consistent with the findings of Kasapidou et al. [61] for milk from ewes raised under an intensive production system.

Milk from ewes in the Control group showed a significantly higher freezing depression point (p < 0.05). The FPD values for milk from the UOP and POP treatments fell within the reported range of −0.547 °C to −0.588 °C [60] while milk from the Control group displayed an even higher FPD. The FPD indicates extraneous water in milk, but since milk from a research farm was used in this study, any intentional water addition that raises the freezing point toward 0 °C is eliminated. There are no studies to our knowledge, on the parameters affecting the FPD of ovine milk. However, in the present study, the difference in FPD between the Control and treated samples is attributed to the diet. Raynal-Ljutovac et al. [62] highlighted several factors that can influence the freezing point of cow’s milk and that are related to the animal feeding, such as diets high in seeds instead of pasture or hay, insufficient roughage, a high protein-to-saturated fat ratio in stalling, and low levels of digestible sugars or energy.

With regard to milk acidity, both active (pH) and titratable, there were no statistically significant (p > 0.05) differences between the treatments. Milk pH was within the reported range (6.51–6.81), but titratable acid was far below the reported range (0.22–0.25% lactic acid) for all treatments [63]. Additionally, milk pH was within the range (6.41–6.79) reported by Pappa et al. [60] in all treatment groups. Milk pH and acidity are widely used as indirect indicators of microbiological quality [61]. Bacterial growth in milk typically involves lactose fermentation, leading to lactic acid production, increased acidity, and a reduction in pH. However, since the milk was sourced from morning milking, stored and transported under refrigerated conditions, and analysed within a few hours after milking, significant bacterial growth and the associated changes in acidity and pH were unlikely to occur.

The electrical conductivity and refractive index values observed were higher than those reported in the review by Park et al. [63] (electrical conductivity: 3.8 mS/cm; refractive index: 1.3492–1.3497) and in the study by Gelasakis et al. [64], which analysed samples from individual animals (electrical conductivity: 4 mS/cm). By contrast, the Brix values were comparable to those reported by Gelasakis et al. [64], specifically 13.8.

Electrical conductivity serves as a quick method for detecting subclinical mastitis in individual animals [65], while the refractive index and Brix value are used to measure total solids in milk [66]. Evaluating the physicochemical properties of milk provides a fast and practical approach for assessing its chemical composition and monitoring animal health at the farm level [64,67].

Yoghurt proximate composition is shown in

Table 5. Yoghurt proximate composition.

Component (%)	Treatment			SEM	Significance
	Control (n = 4)	UOP (n = 4)	POP (n = 4)
Moisture	83.39 b	82.32 a	82.76 a	0.176	**
Ash	0.81 a	0.89 b	0.88 ab	0.020	*
Protein	5.42 ab	5.93 b	5.53 a	0.107	*
Fat	6.06 ab	6.79 b	6.24 a	0.163	*
Carbohydrate	4.32	4.06	4.60	0.153	NS
Total solids	16.61 a	17.68 b	17.24 a	0.176	**
Solids-not-Fat (SnF)	10.55	10.89	11.00	0.160	NS

UOP = unprocessed orange peel; POP = processed orange peel; * = p < 0.05; ** = p < 0.01; NS = non-significant; superscripts a, b differ at p < 0.05.

. There were highly significant differences (p < 0.01) in the moisture and total solids content, while there were significant differences (p < 0.05) in the other examined parameters, i.e., ash, protein, and fat content. By contrast, no significant differences (p > 0.05) were detected in carbohydrate and solids-not-fat (SnF) content across the treatments.

Yoghurt produced from the milk of ewes fed diets containing either unprocessed (UOP) or processed orange peels (POP) had lower moisture content compared to yoghurt from ewes fed the conventional diet (Control). Additionally, ash, protein, and fat content were higher in the yoghurt from ewes on the UOP diet.

Serafeimidou et al. [68] determined the composition of commercially available full-fat sheep milk yoghurt produced by small dairy companies in Greece. Their findings were as follows: moisture content ranged from 77.00 to 86.76 g per 100 g, ash content from 0.87 to 0.98 g per 100 g, protein content from 4.41 to 5.03 g per 100 g, and fat content from 2.31 to 6.89 g per 100 g. Yoghurt proximate composition differed significantly from the composition reported by Pappa et al. [69], that examined commercially available yoghurt of varying types, including traditional and strained yoghurt, as well as low-fat and full-fat varieties. Finally, the composition of commercially available traditional type full-fat yoghurt (12 samples), based on packaging nutrient information, as derived from the websites of main supermarkets that trade in Greece is as follows: protein content ranges from 5.2 to 5.9 g per 100 g, fat content from 6.6 to 6.7 g per 100 g, and carbohydrates 3.2 to 5.0 g per 100 g.

The yoghurt in this study shows notable differences compared to the reported data. Moisture content (82.32–83.39%) is consistent with the findings of Serafeimidou et al. [68] (77.00–86.76%), while yoghurt from the UOP and POP treatments were denser. Protein content in the UOP treatment yoghurt (5.93%) exceeded the upper range reported by Serafeimidou et al. [68] (5.03%) and matched the supermarket data for traditional full-fat yoghurt (5.2–5.9%). Fat content (6.06–6.79%) matches the information provided on supermarket websites (6.6–6.7%) and falls at the higher end of the range reported by Serafeimidou et al. [68] (2.31–6.89%). Ash and carbohydrate content align with previously reported ranges, indicating that the yoghurt from this study, particularly UOP, provide enhanced nutritional profiles compared to traditional options.

Notably, the protein content in yoghurt from ewes on the UOP and POP diets exceeded the 5.5% minimum required by Greek food legislation [28], ensuring compliance with national standards. By contrast, samples from the Control diet had slightly lower protein content. However, protein levels below the 5.5% minimum have also been observed in commercially available yoghurt, both in laboratory-tested samples (4.27–5.03%) [69] and in values reported on product labels (5.2–5.9%). This indicates that deviations from national standards are not uncommon.

Furthermore, the Codex Alimentarius [70] sets a minimum protein requirement of 2.7% and a maximum fat limit of 15% for fermented milk products, confirming that yoghurt from all treatment groups comply with these standards.

3.4. Yoghurt Fatty Acid Composition and Nutritional Indices

The fatty acid composition of the yoghurt is shown in

Table 6. Yoghurt fatty acid composition (% of total identified fatty acids).

Variable	Treatment			SEM	Significance
	Control (n = 4)	UOP (n = 4)	POP (n = 4)
Fatty acid
C4:0	0.93	0.93	0.96	0.014	NS
C6:0	1.60	1.61	1.69	0.025	NS
C8:0	2.27	2.27	2.36	0.036	NS
C10:0	9.25	9.14	9.19	0.089	NS
C12:0	6.43 b	6.16 ab	5.93 a	0.075	*
C13:0	1.23 a	1.12 ab	1.28 b	0.026	*
C14:0	14.11	14.55	13.68	0.216	NS
C14:1	0.62	0.61	0.57	0.018	NS
C15:0	1.27 b	1.02 ab	1.04 a	0.027	***
C15:1	0.31	0.29	0.32	0.007	NS
C16:0	29.98	30.83	30.27	0.388	NS
C16:1	0.27	0.24	0.25	0.007	NS
C17:0	0.63	0.58	0.59	0.012	NS
C17:1	0.29	0.27	0.26	0.006	NS
C18:0	5.93	6.10	6.43	0.171	NS
C18:1 trans	0.42	0.45	0.42	0.022	NS
C18:1 trans-11 (VA)	0.84 b	0.64 ab	0.62 a	0.034	**
C18:1 cis-9	17.39	17.46	18.19	0.343	NS
C18:2 n-6 trans	0.29	0.27	0.24	0.010	NS
C18:2 n-6 cis	4.06	3.80	3.99	0.056	NS
C18:3 n-3	0.99	0.95	0.99	0.047	NS
C18:2 cis-9 trans-11 (CLA)	0.88 b	0.71 ab	0.72a	0.023	**
Lipid class
SFA 1	72.99	73.74	72.84	0.397	NS
MUFA 2	20.15	19.96	20.63	0.331	NS
PUFA 3	6.23	5.73	5.95	0.086	NS
n-3	0.99	0.95	0.99	0.047	NS
n-6	4.36	4.07	4.23	0.057	NS

UOP = unprocessed orange peel; POP = processed orange peel; 1 = saturated fatty acids; 2 = monounsaturated fatty acids; 3 = polyunsaturated fatty acids; * = p < 0.05; ** = p < 0.01; *** = p < 0.001; NS = non-significant; superscripts a, b differ at p < 0.05.

. Caprylic acid (C8:0), capric acid (C10:0), lauric acid (C12:0), myristic acid (C14:0), palmitic acid (C16:0), and stearic acid (C18:0) were the predominant saturated fatty acids, whereas oleic acid (C18:1 cis-9) and linoleic acid (C18:2 n-6 cis) were the major mono-unsaturated and polyunsaturated fatty acids, respectively. With regard to saturated fatty acids, significant differences (p < 0.001–p < 0.05) were found in the levels of lauric acid (C12:0) and the levels of the odd chain fatty acids tridecanoic (C13:0) and pentadecanoic acid (C15:0). In relation to monounsaturated and polyunsaturated fatty acids, significant differences (p < 0.01) were observed in the proportions of vaccenic acid (C18:1 trans-11) and conjugated linoleic acid (C18:2 cis-9 trans-11—CLA), respectively.

The composition of rumen microbiota, influenced by factors such as feed composition and rumen pH, is closely linked to the levels of odd and branched-chain fatty acids in cow milk [71]. Accordingly, differences in the odd-chain fatty acid content between the Control and treated samples can be attributed to variations in rumen pH, which are caused by differences in the diet.

Feeding practices affect the content of CLA in dairy products. Feeding sunflower seed meal to dairy cows resulted in significantly higher levels of CLA in the milk [72]. Additionally, the inclusion of orange peels, either processed or unprocessed, might have affected the rumen biohydrogenation of both linoleic and linolenic (C18:3 n-3) acids affecting the formation of vaccenic acid. The vast majority of CLA is produced in the mammary gland by Δ9—desaturase from vaccenic acid [73].

The observed differences were focused between the yoghurt samples on the Control treatment and the yoghurt on the POP treatment (supplementation with processed orange peels). There was no treatment effect (p > 0.05) in the levels of the lipid classes. Yoghurt fatty acid composition was similar in general to the composition of commercially available sheep milk yoghurt produced by mainly small companies all over Greece [68]. There were few exceptions, such as the lower level of butyric acid (C4:0) or the higher level of CLA, observed in the yoghurt from our study. Additionally, in the present study, the levels of saturated fatty acids (SFA) and polyunsaturated fatty acids (PUFA) were higher than the values reported by Serafeimidou et al. [68] in the previous study. However, Serafeimidou et al. [25] reported a similar fatty acid profile in laboratory produced yoghurt overall. It is important to note that animal diet is the major parameter affecting the fatty acid composition of dairy products.

The nutritional indices of yoghurt fat presented in

Table 7. Yoghurt nutritional indices related to healthy fat intake.

Index	Treatment			SEM	Significance
	Control (n = 4)	UOP (n = 4)	POP (n = 4)
AI 1	3.56	3.74	3.44	0.094	NS
TI 2	3.28	3.45	3.25	0.089	NS
h/H 3	0.47	0.45	0.49	0.014	NS
HPI 4	0.29	0.27	0.29	0.008	NS
PUFA/SFA 5	0.09	0.08	0.08	0.002	NS
DFA 6	32.30	31.79	33.00	0.517	NS

UOP = unprocessed orange peel; POP = processed orange peel; 1 = atherogenicity index; 2 = thrombogenicity index; 3 = hypocholesterolaemic/hypercholesterolaemic ratio; 4 = heath promoting index; 5 = polyunsaturated fatty acid/saturated fatty acid ratio; 6 = desirable fatty acids; NS = non-significant.

are key indicators of fat quality and their implications for healthy fat consumption. Across all indices—atherogenicity index (AI), thrombogenicity index (TI), hypocholesterolaemic–hypercholesterolaemic ratio (h/H), health-promoting index (HPI), polyunsaturated to saturated fatty acid ratio (PUFA/SFA), and desirable fatty acids (DFA)—no significant differences (p > 0.05) were observed among the groups. The AI and TI values were lower, whereas the h/H and DFA values were higher than the ones reported by Paszczyk et al. [74] for sheep milk yoghurt (2.95, 3.04, 0.56, and 37.99, respectively). Finally, the PUFA/SFA ratio was higher in all treatment groups than the value (0.056) reported by Daszkiewicz et al. [75].

The atherogenicity index (AI) and thrombogenicity index (TI), which reflect the potential risk of cardiovascular diseases, remained similar across all treatments, suggesting that incorporating orange peels either unprocessed or processed did not favourably or adversely affect both indices. Both indices exceeded the recommended value, as low values, preferably lower than three, are considered beneficial for human health [61].

The h/H ratio and health-promoting index (HPI), indicators of the balance between beneficial and detrimental fatty acids, also showed no significant variation, indicating that dietary treatments did not influence the balance of cholesterol-modulating fatty acids. The h/H ratio reflects the balance between hypocholesterolaemic and hypercholesterolaemic fatty acids, with higher values being more desirable. Dairy products with a high HPI value are considered healthier for humans, as HPI is the inverse of the atherogenicity index (AI) [37].

The PUFA/SFA ratio, a key indicator of the nutritional quality of fats, was consistently low across all groups, a typical characteristic of dairy products. However, it fell significantly below the Department of Health’s recommended value of 0.45 [76]. A PUFA/SFA ratio above 0.45 is recommended to help prevent cardiovascular and other chronic diseases, as it signifies a healthier fat profile.

The DFA, which represents fatty acids beneficial for human health, was slightly higher in the POP group compared to other treatments, although the difference was not statistically significant. The DFA includes stearic acid and unsaturated fatty acids, which are known for their anti-atherogenic properties, as they help lower plasma cholesterol and triacylglycerol levels [37,77].

3.5. Yoghurt Total Phenolic Content and Antioxidant Profile

Highly significant differences (p < 0.001) were found in yoghurt total phenolic content between the treatments (

Table 8. Yoghurt total phenolic content and antioxidant profile.

Index	Treatment			SEM	Significance
	Control (n = 4)	UOP (n = 4)	POP (n = 4)
TPC 1 (mg GAE/g)	0.038 b	0.035 b	0.026 a	0.012	***
ABTS 2 (μM TE/g)	102.433	96.433	97.767	1.599	NS
DPPH 3 (μM TE/g)	23.492	20.506	22.436	0.661	NS
FRAP 4 (μM TE/g)	66.260	84.027	71.460	6.283	NS

UOP = unprocessed orange peel; POP = processed orange peel; 1 = total phenolic content (TPC); 2 = total antioxidant capacity; 3 = free radical scavenging activity; 4 = ferric reducing antioxidant power; *** = p < 0.001; NS = non-significant; superscripts a, b differ at p < 0.05.

). In general, the total phenolic content of orange peels is affected by the treatment applied for extraction [78]. Yoghurt produced from milk from ewes on the POP treatment had a significantly lower phenolic content than yoghurt from the Control and the UOP treatment. This finding indicates that enzymatic hydrolysis of the orange peels (POP treatment) affected the content of phenolic compounds. On the other hand, the lack of difference between yoghurt from the Control and the UOP treatment shows that replacement of concentrates with orange peels did not negatively affect the total phenolic content of the feed, although differences in the content of individual phenolic compounds might have existed due to the inclusion of the orange peel ingredient.

However, there were no significant differences (p > 0.05) in the total antioxidant capacity (ABTS), free radical scavenging activity (DPPH), and the ferric reducing power activity (FRAP) between the treatments (Table 8), showing that the inclusion of orange peels in the diets of lactating ewes did not result in differences in the antioxidant capacity of the produced yoghurt in relation to samples from ewes on the conventional diet. The antioxidant properties of the samples were assessed using two different types of methods. The Folin–Ciocalteu and the FRAP tests are based on the transfer of one electron, whereas the ABTS and DPPH tests are based on the transfer of both a hydrogen atom and an electron. Furthermore, the Folin–Ciocalteu and the FRAP tests measure hydrophilic antioxidants, while the ABTS test can measure both hydrophilic, and lipophilic antioxidants, and the DPPH hydrophilic antioxidants [79]. Additionally, Chen et al. [80] reported that a thorough understanding of the antioxidant properties of dairy products requires the use of multiple analytical methods, since different components, such as caseins and lipids, have a different sensitivity to different methods. The lack of standardisation among different methods for assessing the antioxidant properties of milk and dairy products was also highlighted by Stobiecka et al. [81] in their recent review. Therefore, the different mechanisms employed by each assay are related to the variations in the expression of the antioxidant activity between the different methods and within samples of the same treatment. It should be noted that the suitability of each assay for determining the antioxidant capacity of dairy products needs further investigation.

The antioxidant capacity of dairy products depends on components such as sulphur amino acids, whey proteins (particularly β-lactoglobulin), vitamins A, E, and C, carotenoids, zinc, selenium, enzyme systems, superoxide dismutase, catalase, glutathione peroxidase, milk oligosaccharides, and peptides that are produced during fermentation. Furthermore, factors such as the presence of various bioactive peptides resulting from the proteolysis of milk proteins by lactic acid bacteria, the fermentation process and post-acidification during storage leading to organic acid production, and the potential aggregation of peptides during the enzymatic hydrolysis of whey protein and casein are all relevant to the antioxidant status of yoghurt [81,82,83].

The total phenolic content of the yoghurt from all treatment groups was lower than the values reported by Terzioğlu and Bakırcı [84] for traditional sheep milk yoghurt. Although these authors also assessed the free radical scavenging capacity (DPPH), their results are not directly comparable due to differences in the assays and measurement units used.

The antioxidant properties of milk and dairy products can generally be improved by incorporating natural additives, either through animal feed or during milk processing. Adding herbal blends or by-products from the fruit and vegetable industry to animal diets enhances the nutritional quality of milk by increasing its content of bioactive compounds and antioxidant potential. However, the antioxidant activity of dairy products is influenced by several factors, including the quality of the raw materials, the type of heat treatment applied, the bacterial starter cultures used, and the inclusion of natural plant additives during processing. These factors may explain the absence of differences in antioxidant status observed between the treatments in the present study [81].

3.6. Yoghurt Active (pH) and Titratable Acidity

The acidity of the yoghurt, both the active acidity (pH) and the titratable acidity (% of lactic acid), is shown in

Table 9. Changes in active pH during storage and titratable acidity.

Storage Day/Variable	Treatment			SEM	Significance
	Control (n = 4)	UOP (n = 4)	POP (n = 4)
Active acidity (pH)
7	4.06 b	4.08	4.08	0.034	NS
14	3.92 a	3.96	3.99	0.026	NS
21	3.88 a	3.90	3.97	0.027	NS
Titratable acidity (g lactic acid/100 g yoghurt)
0	1.38	1.38	1.32	0.044	NS

UOP = unprocessed orange peel; POP = processed orange peel; NS = non-significant; superscripts a, b within the same column differ at p < 0.05

. Titratable acidity was measured on storage day 0, shortly after the yoghurt was removed from the incubation chamber, whereas the changes in active acidity (pH) were monitored during the storage period at weekly intervals, commencing on storage day 7. As shown, there were no differences (p > 0.05) in both the titratable and the active acidity (pH) between the different groups and within the same storage period. The titratable acidity of yoghurt from all treatment groups exceeded the minimum threshold of 0.6% lactic acid recommended by the Codex Alimentarius for commercial products [85]. It is important to note that titratable acidity was measured only on day 0 to assess the potential of growth of pathogens, such as Salmonella spp. and Listeria monocytogenes [57].

As expected, the pH values were reduced during the storage period. The pH values were within the range of 3.69–4.64 reported for commercially available sheep milk yoghurt [69]. Serafeimidou et al. [25] found lower levels of lactic acid content and higher pH values in laboratory produced samples. Parameters such as lipolysis can hydrolyse milk fat into free fatty acids which can lower the pH of the yoghurt. The rate and the extent of lipolysis are also related to the endogenous antioxidants present in the milk [86]. The decrease in pH as well as the increase in lactic acid content during cold storage is linked to post-acidification, which occurs due to the residual activity of microorganisms. These microorganisms continue to produce lactic acid by catabolizing lactose, even during cold storage, gradually enriching the yoghurt with lactic acid [56,87]. Oliveira et al. [88] recommend pH values between 4.0 and 4.4 to prevent an excessively acidic taste in commercial yoghurt. In this study, however, the pH values of yoghurt samples from all treatment groups fell below the minimum recommended level on storage days 14 and 21.

In relation to changes in the active acidity (pH) during storage and within the same treatment, no statistically significant differences (p > 0.05) were observed in yoghurt on the UOP and POP treatments, whereas the pH decline resulted in differences (p < 0.05) in samples from the Control treatment. Serafeimidou et al. [25] reported no differences in the pH values in laboratory produced samples during storage for 15 days.

3.7. Yoghurt Colour

Changes in the colour attributes during storage are shown in

Table 10. Instrumental colour changes during storage (days 7–21).

Storage Day/Variable	Treatment			SEM	Significance
	Control (n = 4)	UOP (n = 4)	POP (n = 4)
Lightness (L*)
7	87.37	85.79	86.30	0.470	NS
14	88.48	87.30	87.41	0.261	NS
21	88.08	87.84	87.04	0.310	NS
Redness (a*)
7	−1.02 b	−1.70 a	−0.80 ab	0.137	*
14	−0.92 ab	−1.59 a	−0.81 b	0.125	*
21	−0.82 ab	−1.65 a	−0.82 b	0.127	**
Yellowness (b*)
7	−10.21 a	−8.23 b	−10.12 a	0.278	**
14	−10.81 a	−8.86 b	−10.49 ab	0.285	**
21	−10.73 a	−8.69 b	−10.19 ab	0.276	**
Chroma (Saturation index)
7	10.31 b	8.60 a	10.17 b	0.236	**
14	10.87 b	9.18 a	10.53 ab	0.245	**
21	10.77 b	9.00 a	10.24 ab	0.238	**
Hue angle
7	252.48	243.73	242.88	6.973	NS
14	242.20	256.46	254.20	6.174	NS
21	265.53	256.42	252.39	4.286	NS
Whiteness index
7	83.52	83.19	82.84	0.379	NS
14	84.10	84.10	83.51	0.176	NS
21	83.87	84.69	83.40	0.243	NS

UOP = unprocessed orange peel; POP = processed orange peel; * = p < 0.05; ** = p < 0.01; NS = non-significant; superscripts a, b differ at p < 0.05.

. As seen, there were no significant differences (p > 0.05) in lightness (L*), hue angle, and the whiteness index, whereas significant differences (p < 0.01–p < 0.05) were found in redness (a*), yellowness (b*), and chroma (colour saturation index) between the treatments. With regard to the redness, yellowness, and chroma values, the differences are mainly observed between the yoghurt on the Control treatment and yoghurt produced from milk from ewes receiving the unprocessed orange peels. Specifically, samples from the UOP treatment had lower redness (a*) and colour saturation index values compared to the Control samples, while their yellowness values were higher than those of the Control samples.

The colour measurements were conducted on the surface of intact yoghurt samples. Since the yoghurt was produced from non-homogenised milk, regarding to the distribution of fat globules, a crust was formed. This crust had a non-even surface that could influence the measurements. The colour of the crust is usually more yellow than the colour of the yoghurt beneath it.

Yoghurt colour from all groups was different than the values reported by Sert et al. [86], who reported lower a* values but higher and positive b* values in sheep milk yoghurt stored for 14 days. Colour parameters are influenced by yoghurt composition. Protein content affect a* values, where a higher protein content results in a greener colour (negative a* values), as shown for yoghurt from the UOP group [46]. However, despite the higher fat content of the UOP group yoghurt, there were no differences in the L* values as expected, since increased fat content raises L* values, as fat contributes to the white appearance of yoghurt [46]. Citrus food residue remaining after juice extraction are a good source of carotenoids [10]. The significant differences observed in yellowness (b*) values might be related to a higher carotenoid content in the UOP feed. The white colour of yoghurt is a typical quality attribute, whereas the whiteness index correlates with the lightness values (L*) [46].

Total colour difference (ΔE_Lab) between the Control and treated samples (UOP and POP) exceeded the value of 1 in all examined periods indicating that minor colour differences could be detected by the human eye (Figure 5). According to Quintanilla et al. [89] the perception of colour difference (ΔE_Lab) depends on both the colour observed and the sensitivity of the human eye. A ΔE_Lab value < 1 indicates that colour differences are undetectable to the human eye, ΔE_Lab values between 1 and 3 represent minor colour differences that can be detected, and ΔE_Lab values > 3 signify colour differences that are easily noticeable to the human eye. The total colour difference values of ΔE_Lab exceeded marginally the value of 1 between samples on the Control and POP treatments on storage days 7 and 14, whereas it was 2.22 on storage day 14. Regarding differences between samples on the Control and UOP treatments the smallest difference, also slightly higher than 1, was observed on day 21.

3.8. Yoghurt Texture and Viscosity

Changes in yoghurt texture, assessed in terms of firmness, cohesiveness, and stickiness, during storage are presented in

Table 11. Changes in yoghurt texture during storage (days 7–21).

Storage day/Variable	Treatment			SEM	Significance
	Control (n = 4)	UOP (n = 4)	POP (n = 4)
Firmness (g)
7	310.50 a	428.75 b	304.00 a	22.509	*
14	310.25 a	517.25 b	330.00 a	30.231	***
21	370.5	452.25	357.75	21.069	NS
Adhesiveness (g f mm)
7	−1825.75	−2578.75	−2119.25	147.132	NS
14	−1935.25 a	−3224.00 b	−2374	219.259	*
21	−2247	−2931	−2419.25	192.386	NS
Stickiness (g)
7	−86	−121.25	−96.75	10.102	NS
14	−92.00 a	−158.75b	−111.75 ab	10.767	*
21	−113.5	−137.75	−113	116.355	NS

UOP = unprocessed orange peel; POP = processed orange peel; * = p < 0.05; *** = p < 0.001; NS = non-significant; superscripts a, b differ at p < 0.05.

. Significant differences (p < 0.01 and p < 0.05) in yoghurt firmness were observed on storage days 7 and 14, respectively. Yoghurt produced from the milk of ewes that received unprocessed orange peel (UOP) was firmer compared to yoghurt from the milk of ewes subjected to the other two treatments. Regarding cohesiveness and stickiness, a treatment effect (p < 0.05) was observed only on storage day 14, with yoghurt from ewes fed the non-hydrolysed orange peel being the most cohesive and stickiest. The textural properties of yoghurt are affected by various parameters, including the quality and composition of the milk (including protein, fat, and total solid contents), the heat treatment of the milk, the milk acidification, and the storage time, among other factors [90]. Abbeddou et al. [16] reported changes in the textural characteristics of yoghurt made from sheep milk produced by animals fed diets containing agro-industrial by-products such as tomato pomace, olive cake or leaves, lentil straw, and Atriplex leaves. Similar to the present study, they found that the differences in texture were neither consistent nor significant enough to adversely affect the quality of the yoghurt.

Finally, in relation to changes in texture during storage within the same group, no statistically significant differences (p > 0.05) were observed in any of the examined parameters across all treatment groups. Domagala [91] and Moschopoulou et al. [92] reported changes in yoghurt textural parameters during storage.

Yoghurt viscosity is shown in

Table 12. Changes in yoghurt viscosity (mm²/s) during storage (days 7–21).

Storage Day/Variable	Treatment			SEM	Significance
	Control (n = 4)	UOP (n = 4)	POP (n = 4)
7	14,497.25	17,801	15,304.00	934.053	NS
14	14,094.5	18,960	15,472.25	1211.679	NS
21	12,444.50 a	21,203.75 b	15,538.25 ab	1424.895	*

UOP = unprocessed orange peel; POP = processed orange peel; * = p < 0.05; NS = non-significant; superscripts a, b differ at p < 0.05.

and, as shown, there were no significant differences (p > 0.05) between treatments on storage days 7 and 14, whereas significant differences (p < 0.05) were observed on storage day 21 between the groups. The rheological properties of the yoghurt significantly affect the product quality. The fat and the protein contents of the milk are the most important parameters. There were fluctuations in the viscosity of the yoghurt within the same treatment. However, according to Moschopoulou et al. [92], the storage period did not affect the viscosity in sheep milk yoghurt throughout storage for 28 days.

In relation to changes in viscosity during storage within the same group, no statistically significant (p > 0.05) changes were observed for all treatment groups.

3.9. Yoghurt Syneresis

The changes in yoghurt syneresis are presented in

Table 13. Changes in yoghurt syneresis during storage (days 7–21).

Storage Day/Variable	Treatment			SEM	Significance
	Control (n = 4)	UOP (n = 4)	POP (n = 4)
Gravity
7	3.21	0.17	0.83	0.897	NS
14	0.42	0.25	0.01	0.117	NS
21	2.48	0.02	0.35	0.824	NS
Centrifugation
7	14.17	11.21	13.83	1.533	NS
14	13.72	12.26	13.49	1.510	NS
21	13.2	9.63	10.82	1.704	NS

UOP = unprocessed orange peel; POP = processed orange peel; NS = non-significant.

. Syneresis during storage was assessed by the drainage (gravity method) and by centrifugation. There were no significant differences (p < 0.05) in yoghurt syneresis between the treatments and within the same storage day with both examination methods. Syneresis, the separation of the liquid phase from the gel, is undesirable in yoghurt and can adversely affect consumer acceptance of the product. Syneresis can happen spontaneously or be triggered by mechanical actions, such as cutting, agitating, or freezing the gel [93]. The texture of yoghurt and its tendency toward syneresis are closely linked to the microstructure of the product, particularly the structure of the protein matrix [91].

Syneresis by centrifugation was different from the values reported by Pappa et al. [69], who examined commercially available yoghurt of various types. This difference is attributed to the lower centrifugation speed applied by the latter researchers and the fact that the values in this study were derived from different types of yoghurt, such as strained, traditional, full-fat, or low-fat.

Syneresis decreased during storage in yoghurt from all treatments. This finding is in agreement with the findings of Domagała [91]. Regarding syneresis by gravity, Domagała reported low syneresis in yoghurt from sheep milk. The total solid and the protein content are related to the extent of syneresis. Although there were differences in the total solid content between the treatments (Table 5), this effect was not born out in the syneresis extent during storage. This effect is attributed to the fact that there were no differences in the protein content between the treatments. According to Domagała [91], the protein matrix of sheep milk yoghurt produces a stronger gel that is more resistant to deformation.

The extent of syneresis, determined either by the drainage or the centrifugation method, was not significantly (p > 0.05) affected during storage within the same treatment group.

3.10. Yoghurt Sensory Evaluation

The taste panel evaluation of the yoghurt is shown in Figure 6. There were no significant differences (p > 0.05) between the treatments in any of the examined sensory traits. Except for acidity and syneresis, samples from all treatments scored above the acceptability threshold (score = 4) for positively perceived sensory characteristics. Acidity values exceeded the acceptability threshold, indicating that the yoghurt had an acidic taste. However, based on the pH values measured on day 7 of storage (Table 9), when the sensory evaluation was conducted, the acidic taste was not overly intense, as the pH values remained within the recommended range of 4.0–4.4, associated with a moderate acidic taste [88]. An astringent flavour and bitter taste have been reported in yoghurt with the direct addition of grape seed extract at a 1% polyphenol content [94], indicating that higher polyphenol levels can negatively impact yoghurt acceptability. Regarding acidity, the higher polyphenol content in samples from the Control and UOP groups (Table 8) may explain the observed differences. Regarding syneresis, an undesirable characteristic in yoghurt, it was below the acceptability threshold, indicating that the separation of the liquid phase during spoon cutting and after 7 days of storage was rather low.

As observed in the organoleptic analysis, despite significant differences in proximate composition, total phenolic content, and firmness, these effects were not reflected in the related sensory characteristics or overall acceptability. This finding may be attributed to the fact that the panellists were untrained, and the panel size was relatively small, unlike similar studies where a larger number of trained panellists evaluated the products [95,96,97].

Unlike the present study, the use of agro-industrial by-products, such as tomato pomace, olive cake or leaves, lentil straw, and Atriplex leaves, has resulted in significant differences in texture, taste, and odour, despite no off-flavours being reported. However, these differences were not deemed prohibitive for incorporating these by-products into sheep diets [16].

4. Conclusions

The present study demonstrated that supplementing ewes’ diets with either processed or unprocessed orange peels has no adverse effects on the quality characteristics of the resulting yoghurt compared to those made from milk produced by ewes fed a conventional diet. Notably, the inclusion of unprocessed orange peels (UOP) resulted in yoghurt with higher protein and fat content than those from ewes fed processed orange peels (POP), while fat nutritional indices remained consistent across all treatments. Antioxidant activity differed only in the POP group when measured as total phenolic content. Differences were observed in colour parameters, such as redness (a*), yellowness (b*), and colour saturation (Chroma); however, Lightness (L*) and the whiteness index—key indicators of consumer perception—remained unaffected. Textural analysis revealed that yoghurt from ewes fed UOP was firmer, more cohesive, and stickier during the first two weeks of storage. Taste panel evaluations showed no significant differences in flavour across treatments, with all yoghurt samples receiving an overall acceptability score above 4. Instrumental assessments of sensory attributes, such as colour, texture, viscosity, and syneresis, confirmed stability throughout storage.

These findings highlight the practical potential of incorporating orange peels—both processed and unprocessed—as a sustainable secondary feed ingredient in dairy production. Their use could reduce feed costs, enhance milk composition, and contribute to circular economy practices in the dairy industry. However, certain limitations of this study should be acknowledged. As this research was conducted under controlled experimental conditions, large-scale trials in commercial dairy farms would help validate the practical feasibility of incorporating orange peels into ewe diets. Future studies should explore both their suitability within various feed formulations and their effects on other dairy products, such as cheese and butter, to expand their potential applications in the dairy industry.