per-Effects of polymerized goat milk whey protein on physicochemical properties and microstructure of recombined goat milk yogurt

Goat milk whey protein concentrates were manufactured by microfiltration (MF) and ultrafiltration (UF). When MF retentate blended with cream, which could be used as a starting material in yogurt making. The objective of this study was to prepare goat milk whey protein concentrates by membrane separation technology and to investigate the effects of polymerized goat milk whey protein (PGWP) on the physicochemical properties and microstructure of recombined goat milk yogurt. A 3-stage MF study was conducted to separate whey protein from casein in skim milk with 0.1-µm ceramic membrane. The MF permeate was ultrafiltered using a 10 kDa cut-off membrane to 10-fold, followed by 3 step diafiltration. The ultrafiltration-diafiltration-treated whey was electrodialyzed to remove 85% of salt, and to obtain goat milk whey protein concentrates with 80.99% protein content (wt/wt, dry basis). Re-combined goat milk yogurt was prepared by mixing cream and MF retentate, and PGWP was used as main thickening agent. Compared with the recombined goat milk yogurt without PGWP, the yogurt with 0.50% PGWP had desirable viscosity and low level of syn-eresis. There was no significant difference in chemical composition and pH between the recombined goat milk yogurt with PGWP and control (without PGWP). Vis-cosity of all the yogurt samples decreased during the study. There was a slight but not significant decrease in pH during storage. Bifidobacterium and Lactobacillus acidophilus in yogurt samples remained above 10 6 cfu/g during 8-wk storage. Scanning electron microscopy of the recombined goat milk yogurt with PGWP displayed a compact protein network. Results indicated that PGWP prepared directly from raw milk may be a novel protein-based thickening agent for authentic goat milk yogurt making.


INTRODUCTION
Fermented goat milk products are considered as specialty foods due to their nutritional and functional attributes, and their market share is increasing (Guo, 2020).Unlike cow milk yogurt, goat milk yogurt has a weak body with low viscous state and low water-holding capacity, which may be due to the difference in casein content and seasonal changes in chemical composition (Li and Guo, 2006).The most common methods to improve the weak texture of yogurt by increasing the total solids and adding thickening agents such as pectin (Herrero and Requena, 2006;Wherry et al., 2019).Whey protein products have many functional properties in food systems including gelation, thickening, and water-holding capacity (Bryant and Mcclements, 1998), and also be used to increase the total solids of yogurt.Wang et al. (2012) achieved an acceptable goat milk yogurt with the addition of whey protein isolate.
There has been a great interest in the application of membrane separation technology for separation and recovery of milk components in the dairy industry (Cancino et al., 2006).Microfiltration (MF) with a membrane pore size of 0.1 to 0.2 µm is commonly used to retain casein in the retentate, and whey protein passes through the membrane (Reale et al., 2020).For this reason, MF is a useful tool for the separation of whey protein from casein, and resulting in a casein-rich retentate fraction and a whey protein-rich permeate fraction (Adams and Barbano, 2013).The MF retentate could be used in the formulation of different dairy products due to its nutritional and functional values and its stability during processing (Tremblay-Marchand et al., 2016).It is possible to recombine MF retentate and cream to prepare cheese (Govindasamy-Lucey et al., 2007;Lu et al., 2016).Whey protein in the MF permeate is nutritionally and functionally superior to whey protein from cheese whey, which gives it a per-fect potential for preparing whey protein concentrate by a UF system (Saboyainsta and Maubois, 2000).Ultrafiltration was used to concentrate protein, due to the retention of protein and selective permeation of lactose, minerals, water, and compounds of low molar mass (Carter et al., 2021).Baldasso et al. (2011) used UF technology to concentrate protein.When the feed concentration reaches a certain level, diafiltration (DF) is introduced in the process to allow for greater purification of protein and to overcome the high rise in viscosity (Gavazzi-April et al., 2018).Baldasso et al. (2011) showed that the addition of small DF volumes several times is more effective than a big volume added one time only.The UF-DF combinations were used to concentrate and purify goat milk whey protein, and to improve the degree of protein purification.
Caprine milk whey protein concentrate can be obtained by membrane separation technology, which has good gelation property (Sanmartín et al., 2015).Heat treatment could be used to prepare soluble whey protein aggregates.When heated at certain temperature and protein concentration that would increase the aggregate size which further increase the gel strength.Heat-induced polymerized whey protein is regarded as a promising thickening agent to improve goat milk yogurt structure (Wang et al., 2017).According to Bierzuńska et al. (2019), one of the ways to improve firmness and syneresis of yogurt is the addition of polymerized whey protein.Fang and Guo (2019) made it possible to incorporate polymerized liquid whey protein concentrate produced directly from cheese whey by membrane separation technology in yogurt production.Therefore, we hypothesized that polymerized goat milk whey protein (PGWP) prepared directly from the raw milk may be a more effective thickening agent for recombined goat milk yogurt production.
The objective of this study was to prepare goat milk whey protein concentrates by MF and UF associated with DF and to develop a recombined goat milk yogurt by recombining cream with MF retentate of skim milk.The polymerized goat milk whey protein was used as a thickening agent, and the physicochemical properties, microstructure, and shelf life stability of the recombined goat milk yogurt were investigated.

Membrane Processing
Microfiltration.Raw goat milk was heated to 55°C and skimmed with a separator (SA 10-T, Frautech SRL) to obtain skimmed goat milk and cream.The fat contents (± SD) in skimmed goat milk and cream were 0.13 ± 0.01% and 42.0 ± 0.32%, respectively.The cream was placed at refrigerator (4°C), and the skimmed goat milk was microfiltered to approximately a 3× volumetric concentration factor (VCF).
A laboratory-scale MF system (APPM-C-50, Gaochen Environmental Protection Equipment Co. Ltd.) equipped with ceramic membranes (nominal pore diameter = 0.10 µm; surface area = 0.25 m 2 ) was used.Ten tubular ceramic membranes measuring 1.0 m in length were housed in a tubular stainless steel module.The MF process was operated at a transmembrane pressure of 0.9 bar with the temperature of 50°C.When a VCF of 3× had been reached, then the second stage (DF step) followed, during which deionized water at 50°C was added into the feed.The volume of permeate collected in first stage corresponded to the volume of added water.For the second stage, the target VCF was 3×; when a 3× VCF was achieved, the DF step was repeated to complete a 3-stage process.The MF system was stopped when the MF 3-stage process achieved.The MF retentate was stored at 4°C, the permeate from the 3-stage was mixed together and used as the feed solution for the following UF process followed by 4 DF (DF1-DF4).Both MF and UF ran in the retentate recycling mode by a centrifugal feed pump.
Ultrafiltration of MF Permeate.The UF system (UF-10KD, Gaochen Environmental Protection Equipment Co. Ltd.) was carried out using one cassette of spiral-wound membrane with a molecular weight cutoff of 10 kDa.The UF membrane can retain protein in the feed liquid and it allows selective permeation of low molar mass compounds such as lactose, minerals, and water.The UF process was performed at a transmembrane pressure of 1.1 bar with the temperature of 45°C.The UF was operated by continuously removing the permeate stream until the desired VCF was achieved.Then 4 DF (DF1-DF4) followed, which was to add some incremental volumes of demineralized water to concentrate, during which the added volume was removed.The amount of water added corresponded to the amount of UF retentate.The permeate flux J (L m −2 h −1 ) was measured during MF and UF processes every 10 min and calculated according to the following equation: where Vp (L) is the amount of permeate collected during the period t (h) given the permeation surface area of membrane S (m 2 ).
Electrodialysis.The goat milk whey protein concentrates were demineralized by electrodialysis (ED).The ED operation was carried out using an FTED-40 module prepared by FuMA-Tech GmbH.The cation and anion exchange membranes were provided by Astom Co., Ltd.The conditions used for goat milk whey protein desalination were as follows: laboratory temperature (45 ± 2°C), constant voltage (20.0 ± 0.2 V), electrolyte: anhydrous sodium sulfate (10 g/L), concentrate: distilled water acidified with HNO 3 (pH 2).The ED process was stopped as the conductivity was reached (0.12-0.15 mS/cm).Conductivity and pH were measured with a conductivity meter (Crison) and pH meter (Mettler Toledo Ltd.) during ED process.Subsequently, the desalted goat milk whey protein concentrates were freeze-dried in a freeze dryer (Alpha 1-2, Marin Christ Inc.) to obtain goat milk whey protein powder for compositional analysis.The cleaning of the fouled membrane was carried out by using a cleaningin-place procedure at the end of each experiment.
Chemical Analyses and Lactose Separation.Kjeldahl method was used to determine total nitrogen (TN), noncasein nitrogen (NCN), and NPN (Marshall, 1993).True protein (TP) was calculated as TP = (TN − NPN) × 6.38, CN = (TN − NCN) × 6.38, and whey protein = (NCN − NPN) × 6.38.Total solids and ash content were determined as described by (Marshall, 1993).Lactose was determined by the dinitrosalicylic acid method (Miller, 1959).All experiments were performed on 3 trials, and each trial was carried out in triplicates.The lactose removal percentage was calculated by dividing the mass of lactose in permeate by the mass of lactose in the initial feed volume and multiplying by 100.
Preparation of PGWP.The goat milk whey protein powder was dissolved in deionized water to obtain a 10% (wt/vol) its whey protein solution and stored at 4°C for 12 h to complete hydration.The solution was adjusted to pH 7.7 with 1 M sodium hydroxide, heated to 75°C for 25 min with continuous stirring, and was then quickly cooled to room temperature and marked as PGWP.
Yogurt Sample Preparation.Yogurt bases were prepared by mixing cream with MF retentate, and the ratio of cream to MF retentate at 10:90.The premixed sugar (6%, wt/vol) and pectin (0.3%, wt/vol) were added to the yogurt bases.The mixed blends were homogenized at 200 bar at 55°C and subsequently heattreated at 85°C for 15 min in a water bath.After heat treatment, the mixtures were rapidly cooled to 43°C for the addition of ABY-3 starter culture (0.03%, wt/ vol) and PGWP solution.All mixtures were transferred to yogurt cups and incubated at 43°C for 4.5 h, and then stored at 4°C until analyzed.The inoculated mix was filled in sterile glass cups (150 mL) covered with lids.The PGWP was added separately at a content of 0.10, 0.20, 0.30, 0.40, 0.50, and 0.60% (wt/vol) and the yogurt without PGWP was chosen as a control.
Viscosity Analysis.Viscosity was measured by a Brookfield viscometer (DV-3, Brookfield Engineering Laboratories Inc.).Yogurt samples were brought back up to room temperature for 2 h and stirred in the same direction for 30 s. Measurements were set to 200 rpm for 30 s with a LV4 spindle (Wang et al., 2015).Each experiment was performed in triplicate for 3 trials.
Syneresis Test.Syneresis was measured according to Li and Guo (2006).The recombined samples (30 g) were fermented in centrifuge cups for 4.5 h and stored at 4°C.The samples were centrifuged (Avanti J-E, Beckman Coulter, Brea, CA) at 640 × g for 10 min at 4°C.Supernatant liquid was removed and weighed.The syneresis was calculated as the percentage of whey weight to yogurt weight.All experiments were carried out for 3 trials, and each trial was conducted in triplicate.
Physicochemical and pH Analysis.The TS content was measured with a moisture meter (MJ33, Mettler Toledo).The protein, fat, and ash contents were determined according to methods 991.23, 989.05, and 945.46, respectively, of AOAC International (2002).Protein content was measured by Kjeldahl method with a conversion factor of 6.38.Fat content was assayed with Gerber centrifugation (110 × g, 5 min, 60°C).Ash content was determined by dry-ashing with a muffle furnace (SX-2.5-12,Jingke).Carbohydrate content was calculated by TS minus protein content, fat content and ash content as described by Wang et al. (2015).The pH values were measured with a pH meter (PHS-3C, Jingke).All experiments were carried out in triplicate for 3 trials.
Shelf Life Test and Survivability of Probiotics.During storage, viscosity, pH, and microbiological analyses were performed weekly for 8 wk.The survivability of probiotics was performed according to the method of Wang et al. (2012) with some modifications.Lactobacillus acidophilus was cultivated on MRS agar with maltose.Bifidobacterium was selectively grown on MRS agar with glucose and dicloxacillin, lithium chloride, and l-cysteine hydrochloride were further added to the agar.Lactobacillus acidophilus and Bifidobacterium were anaerobically incubated at 37°C for 3 d.Coliform, mold, and yeast counts were determined by using Petrifilm plates (3M Petrifilm).Coliform plates were stored at 35°C for 24 h.Mold and yeast plates were incubated at 21°C for 5 d.
Microstructure Examination.Microstructure of the yogurt samples were examined by scanning electron microscopy according to the procedures of Wang et al. (2017) with modifications.The samples were fermented in agar cubes and stored at 4°C for 12 h.The agar cubes were fixed in 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.2) for 24 h.The samples were then washed in the same buffer for 3 times and fixed in 1% osmium tetroxide followed by 3 rinses in diluted (50 mM) cacodylate buffer (pH 7.2).A series of ethanol solution were used for dehydration.The dehydrated samples were dried using a freezer dryer (Alpha1-2, Marin Christ Inc.).Samples were mounted on SEM stubs and sputter coated with 3 nm of Au/Pd (80/20) alloy and examined by SEM (JSM-6700F, Jeol Ltd.).

Statistical Analysis.
All experiments were carried out in triplicate.The results of triplicate experiments were statistically analyzed and expressed as mean ± standard deviation.The ANOVA (P < 0.05) and Tukey's test were calculated using SPSS 20 software (IBM Corp.).All figures were drawn using Origin 8.0 (OriginLab Corp.).

Membrane Separation
Permeate Flux of MF Process.Permeate flux of a 3-stage MF process was measured with processing time.The permeate flux of each stage of the 3-stage decreased during processing and finally tended to be constant (Figure 1).The results were similar to that of Adams and Barbano (2013) and Benedetti et al. (2015).This phenomenon may be associated with an increase in the thickness of the gel layer through the membrane surface leading to an increase in resistance to solvent flow, and resulting in a decline in permeate flux (Benedetti et al., 2015).The permeate flux increased from the first to the second stage, and a similar trend was observed between the second and third stage (Figure 1).The increases in permeate flux in stages 2 and 3 were due to the addition of deionized water as a diluent for the DF stages and thus resulted in a decrease in feed viscosity (Adams and Barbano, 2013).
Chemical Composition.The TS, NPN, NCN, and whey protein values in MF retentate decreased (P < 0.05) with successive processing stage (Table 1), which was due to diafiltration using water as a diluent after stage 1 and 2, and the passage of lactose, whey protein, and low molecular compounds through the membrane.The TN, TP, and CN values increased with increasing stages.We expected that TN and TP values would decrease in MF retentate, as whey protein passed through the membrane to the permeate with the MF stages increased.The discrepancies may be the result of slight changes in VCF between the second and third stages.The similar changes were also observed by Adams and Barbano (2013) that slight variations in VCF were discovered during replicates.
Permeate TS, TN, NPN, NCN, TP, CN, and whey protein values decreased (P < 0.05) with increasing stage (Table 2).This was mainly due to the MF retentate was diafiltrated using deionized water as a diluent.The finding was in agreement with that obtained by Adams and Barbano (2013) who have revealed that a 3-stage, 3×, feed-and-bleed MF was conducted to remove whey protein from skim milk, and the TS, TN, NPN, NCN,TP, and whey protein content of permeate decreased with each stage.that the permeate flux decreased with the increasing of VCF.The behavior could be linked to the increase of the osmotic pressure of the retentate.As reported by Bacchin et al. (2006) that the permeate flux decreased with increasing concentration of protein, which may be due to the more accumulation of solute molecules in the polarized layer, resulting in an increase in permeation resistance.The permeate flux did not show much variation when the VCF of UF processing reached a maximum of 11.The largest and most dense layer was formed at higher VCF, which decreased the permeate flux until it reached a static condition (Atra et al., 2005).Diafiltration processing was performed when VCF reached 10 and the volume was suitable for that purpose.The DF stages were investigated to obtain goat milk whey protein concentrates with high protein content.
From Figure 3 it can be seen that the percentages of protein on a dry basis (protein content/TS content) versus the stage of the process increased with increasing DF stages.The protein content in retentate reached 67.4% at the end of the UF processing, and significantly (P < 0.05) increased from DF1 to DF3.The increased protein content was attributed to a reduction of the retentate lactose and minerals were gradually eluted by adding water (Baldasso et al., 2011).The lactose removal rate increased with increasing DF stages and reached 82.5% after DF4 (Figure 3).The results were similar to the literature data as discovered by Butylina et al. (2006) that the compounds of low molecular weight (lactose and minerals) through the UF membrane, which increased the lactose removal rate of the retentate.The result suggested that UF in association with DF mode is a promising method and can be used to concentrate and purify the goat milk whey protein.
Electrodialysis.The conductivity of goat milk whey protein concentrates decreased as a function of ED process (Figure 4).The conductivity curve had a rapidly decline in the beginning 60 min, when ion was rapidly transported.The conductivity showed slowly loss and reached 0.38 mS/cm in the end, which was due to a reduction of the process driving force because the ions concentration in the solution were gradually decreased (Diblíková et al., 2013).From Figure 4, it can be seen that more than 85% of all salts were removed by 100 min of ED process.As the selected conductivity was reached, the composition of desalted goat milk whey protein concentrates was 80.99% protein content, 18.67% lactose content, and 0.34% ash content (wt/wt, dry basis).
Preliminary Results.The viscosity and syneresis of the yogurt samples with different levels of PGWP were shown in Figure 5.The results showed that the viscosity of yogurt with the addition of PGWP was significantly higher (P < 0.05) than that of the control sample without PGWP (Figure 5A).This may be due to the interactions between PGWP and casein micelle to form a complex system, resulting in an increase in viscosity of the yogurt with PGWP (Cheng et al., 2017).The viscosity of the yogurt samples increased Values in the same column not sharing a common superscript are significantly different (P < 0.05).
The yogurt fortified with 0.60% PGWP showed the highest viscosity, and there was no significant difference between samples with 0.50 and 0.60% (P > 0.05).This result was similar to that obtained by Bierzuńska et al. (2019) who have revealed that incorporation of polymerized whey protein can increase the viscosity of yogurt.
Syneresis refers to the shrinkage of the gel, leading to whey separation.Whey separation was defined as the appearance of whey on the gel surface of fermented dairy products (Hossain et al., 2020).All samples were not observed the behavior of whey separation during postacidification after overnight.The yogurt with the addition of PGWP showed lower (P < 0.05) syneresis values than control sample (Figure 5B).As whey protein gel has water-retaining capacity, the recombined goat milk yogurt prepared by the addition of PGWP has the ability to maintain the stability of the water phase in the yogurt network (Gilbert and Turgeon, 2021).The results suggested that the yogurt with the combination of PGWP showed better water-holding capacity.Yogurts with PGWP levels ranging from 0.00 to 0.50% had significant effect on syneresis (P < 0.05).When 0.60% PGWP was added to the yogurt the syneresis was the least, and there was no significant difference in syneresis between PGWP levels of 0.50 and 0.60% (P > 0.05).Our findings were similar with the results of Wang et al. (2017) showing that the addition of polymerized whey protein to goat milk yogurt decreased syneresis of yogurt.
Physicochemical Properties.The composition and pH of the control and the yogurt with PGWP were listed in Table 3.There were no significant differences (P > 0.05) in chemical composition between the control and the yogurt with different levels of PGWP.As reported by Li and Guo (2006), the chemical composition of yogurt was affected by the type of raw materials used, type of yogurt manufactured, and fortification methods.The pH values were similar and no significant differences (P > 0.05) were observed between the control and the yogurt with different levels of PGWP, which suggested that the yogurt with the addition of   PGWP did not significantly influence the activity of yogurt starter cultures.
Shelf Life Test and Survivability of Probiotics.Changes in pH and viscosity of the control and the yogurt with PGWP were shown in Figure 6.There was no considerable decrease in pH for the control and the yogurt with PGWP during 8-wk storage (Figure 6A).The slight decrease in pH was probably due to the production of lactic acid by starter cultures during storage (Shah et al., 1995).The results were similar with the findings of Wang et al. (2012) who reported that the pH of goats' milk yogurts decreased upon 12-wk storage.It can be seen that different PGWP addition levels did not significance change pH values of yogurts within the same storage time, which indicated that pH of the yogurt samples was less likely to be influenced by adding PGWP.
Figure 6B showed changes in viscosity of the control and the yogurt with PGWP.Viscosity was initially higher for the yogurt with the addition of PGWP than the control sample.There was a gradual decrease in viscosity of the control sample and the yogurt with PGWP during storage.The proteases in milk may continue to decompose protein in the yogurt network after fermentation, resulting in loosening of the gel network and leading to a decrease in viscosity (Wang et al., 2015).The viscosity of the yogurt samples increased with PGWP added increasing from 0.00 to 0.60% during the same storage time, and there was no significant difference between samples with level ranged from 0.50 to 0.60% except for wk 7 (P > 0.05).This characteristic could be responsible for the difference in TS content (Wang et al., 2012).
Microbiological analyses were shown in Figure 7 during storage.The initial count for Bifidobacterium in the control sample and the yogurt with PGWP was more than 10 9 cfu/g (Figure 7A).Counts of the Bifidobacterium gradually decreased for both the 2 samples  during 8-wk storage, but was still remained above 10 7 cfu/g over the 8-wk shelf life.The population of L. acidophilus was above 10 7 cfu/g for the initial 4 wk in the control sample and the yogurt with PGWP (Figure 7B).Lactobacillus acidophilus showed a gradual decline and was still well above 10 6 cfu/g during 8-wk storage.The decrease in L. acidophilus population may be due to the formation of hydrogen peroxide by L. bulgaricus during storage, which might inhibit the growth of L. acidophilus (Zhang et al., 2015).
There was no significant difference in the population of Bifidobacterium and L. acidophilus within the same storage time among the yogurt samples produced by different PGWP contents (P > 0.05), indicating that PGWP did not have an effect on the viability of the Bifidobacterium and L. acidophilus during storage.Compared with L. acidophilus, Bifidobacterium showed a better viability during storage.The possible explanation may be that S. thermophilus in the yogurt was used as an oxygen scavenger, creating an anaerobic environment that was beneficial for survival of Bifidobacterium (Lourens-Hattingh and Viljoen, 2001).There was no coliform, mold, or yeast detected in the 2 yogurt samples during storage, and the results indicated that the yogurt samples were safe and clean after storage at 4°C for 8 wk.
Microstructure.The scanning electron micrographs of the control and the yogurt with 0.50% PGWP were shown in Figure 8.The addition of PGWP showed clear variation in the degree of networking in the micrographs.Compared with the control yogurt (Figure 8A and 8C), the yogurt with 0.50% PGWP showed denser and well-organized protein clusters with high connectivity network structure (Figure 8B and 8D, the area was indicated by the white arrow).When the milk system was inoculated with starter cultures, the PGWP might interact with casein micelle to form bridges during acidification (Schorsch et al., 2001), which improved the water-holding capacity of the yogurt with PGWP (Vukic et al., 2014).

CONCLUSIONS
The goat milk whey protein concentrates with high protein content (80.99%) can be manufactured by microfiltration followed by ultrafiltration associated with diafiltration, and microfiltration retentate and the cream separated from the initial milk were recombined to make yogurt.The consistency and syneresis of the recombined goat milk yogurt were improved with the addition of PGWP.The survival rates for Bifidobacterium and L. acidophilus were stable and remained 10 6 cfu/g during the 8 wk of storage.Microstructure analysis indicated a denser protein network for the sample fortified with PGWP.The results of this study demonstrated that PGWP prepared from raw milk by membrane separation technology could be used as cothickening agent for pure goat yogurt making.

Figure 1 .
Figure 1.Permeate flux during processing time for 3-stage at 50°C with 0.10-µm ceramic membrane using 2 diafiltration (DF) stages.Error bars represent the SD of the means (n = 9).
Figure 2. Microfiltration permeate concentration by ultrafiltration: influence of volumetric concentration factor (VCF) on permeate flux.Error bars represent the SD of the means (n = 9).

Figure 3 .
Figure 3. Percentage of protein content and lactose removal rate (wt/wt, dry basis) versus the stage of ultrafiltration (UF) and diafiltration (DF) processes.Error bars represent the SD of the means (n = 9).Different letters (a-d) denote significant differences (P < 0.05) among samples.

Figure 4 .
Figure 4. Conductivity and demineralization rate changes during demineralization of goat milk whey protein concentrates.Error bars represent the SD of the means (n = 9).
Figure 5. Effects of polymerized goat milk whey protein (PGWP) level on viscosity (A) and syneresis (B) of recombined goat milk yogurt.Error bars represent the SD of the means (n = 9).Different letters (a-f) denote significant differences (P < 0.05) among samples.

Figure 6 .Figure 7 .
Figure 6.Effects of polymerized goat milk whey protein (PGWP) level on pH (A) and viscosity (B) of recombined goat milk yogurt during storage.Different letters (a-g) within the same period indicate significant differences (P < 0.05) among samples.Error bars represent the SD of the means (n = 9).

Figure 8 .
Figure 8. Scanning electron microscopy photographs of the control, yogurt without polymerized goat milk whey protein (A and C), and the yogurt with 0.50% polymerized goat milk whey protein (B and D).The white arrows show the connectivity network structure of the yogurt samples.Photos A and B are 2,000× and B and D are 5,000×.
Tian et al.: POLYMERIZED GOAT MILK WHEY PROTEIN AS THICKENING AGENT

Table 1 .
Tian et al.:POLYMERIZED GOAT MILK WHEY PROTEIN AS THICKENING AGENT The composition (% by weight) of retentate from each stage of the 3-stage microfiltration (MF) process1,2

Table 3 .
Chemical composition and pH of yogurt samples 1