3 experimental part

Determination of a rational amount of starter and rennet for the production of soft cheese from a mixture of raw milk

Since cow's and goat's milk differ in their technological properties, which is primarily due to the fractional composition of the proteins, it seemed appropriate to specify the parameters of the technological process for the production of soft cheese from a mixture of dairy raw materials.
The samples are based on the classic soft cheese technology of Moale cheese, produced with the use of rennet
curdling, and Tea-Cheese technology when using Acid-Cheese curdling.
When working out the product production parameters the amounts of rennet and bacterial starter applied to soft cheeses from a mixture of milk were determined. In the production of soft cheeses from a mixture of raw milk rennet and direct application bacterial starter with a specific strain combination were used.
The research was carried out in two phases. A series of all experiments were carried out under the same conditions: at (28-30) °C; the clot formation time for rennet clotting was minutes90 and for acid rennet clotting 360 minutes.
In the first series of trials the amount of yeast was kept constant, the amount of rennet added varied between g/1003 kg and 7 g/100 kg in steps of 2. The amount of calcium chloride added in all experiments was 40 g dry salt per 100 kg of milk. In the production of soft cheese samples, the quality parameters of clots obtained by clotting goat's milk, cow's milk and their mixed compositions with different doses of enzyme were studied. A milk clotting enzyme with standard activity was selected for a series of experiments. The results of the experiment are shown in figure 3.
The analysis of the mathematical models of the dependencies shown in the figures shows that the enzyme preparation exhibits a greater rate of action in a milk mixture dominated by cow's milk. However, it should be noted that as the amount of enzyme preparation in the milk mixture decreases, its speed of action also decreases, which may be due to the difference in the fractional composition of the milk protein mixture, namely, a higher content of β-casein in goat milk.

a - rennet curdling

b - Acid-rich curdling

Figure 3 - Change in active acidity during milk coagulation with different amounts of rennet
▲ - 3 g/100 kg rennet

• 
  - 5 g/100 kg rennet x - g/1007 kg rennet
  

On the basis of the research the optimum dose of rennet for soft cheese mixture in the amount of 5 g/100 kg was selected, which allows to produce soft cheese with the required quality indicators. Increasing the amount of injected enzyme in this case led to thickening of the clot structure and, as a consequence, increasing its syneretic properties.

In the second series of experiments, the amount of enzyme preparation introduced was taken as a constant value. The quality parameters of clots obtained by clotting goat's milk, cow's milk and their mixed compositions were studied with different amounts of introduced bacterial leaven, which varied in the range from up to 1%4 with step The results1,5. of experiments are shown in Figure 4.

a - rennet curdling

b - Acid-rich curdling

Figure 4 - Change in titratable acidity during milk coagulation with different amounts of bacterial starter

• 
  - 1% bacterial sourdough
  

• 
  - 2.5% bacterial sourdough
  

▲ - 4% bacterial sourdough
As can be seen from the figure, changing the dose of bacterial starter in the mixture had a significant effect on the clotting time and the titratable acidity of the clot.
The acidity of the clots at the end of rennet curdling is (65±2)°T, for acid-rennet curdling (70±2)°T. It was found that titratable acidity exceeding 75° T for either type of clotting was overestimated, resulting in an excessively acidic taste.
It should be noted that increasing the amount of clotting enzyme did not in all cases have a positive effect on the quality of the clots and their ability to retain moisture. When the dose of enzyme preparation in the milk mixture was increased, a change in the properties of the cheese clot was observed, and the waste of milk constituents into the milk whey, which affected the physico-chemical properties and colour of the whey.
With an increase in the dosage of added yeast there was a thickening of the consistency, an increase in whey secretion and excessive acidity, which has a negative effect on the organoleptic qualities of the product. Tables 17, 18 show the whey solids content of the soft cheese samples with different r.p.d. of rennet and bacterial starter.
Table 17 - Dependence of m.d.s. and whey yield on the amount of rennet

Indicator	Amount of rennet, g/100 kg
	3	5	7
	rennet curdling
Mass fraction of solids, %	4,6±0,11	4,9±0,14	5,3±0,13
Serum yield, %	56±10	76±9	82±9
	acidic clotting
Mass fraction of solids, %	5,3±0,17	5,5±0,18	6,7±0,19
Serum yield, %	59±11	80±8	86±7

Table18 - Dependence of m.d.s. and whey yield on the amount of bacterial starter

Indicator	Amount of bacterial inoculum, %
	1	2,5	4
	rennet curdling
Mass fraction of solids, %	4,8±0,15	5,1±0,13	5,5±0,17
Serum yield, %	58±6	70±8	76±7
	acidic clotting
Mass fraction of solids, %	5,1±0,19	5,4±0,17	5,9±0,19
Serum yield, %	62±5	78±7	88±7

The addition of 2.5% starter to a 1:1 mixture of cow's and goat's milk contributed to a soft cheese with good organoleptic characteristics - clean, sour milk flavour and a delicate homogeneous consistency without whey separation.
The analysis of parameters of pilot samples allows to recommend to use in the technology of soft cheese from a mixture of cow's and goat's milk with 2.5% starter and 5 g/100 kg rennet, as exactly these quantities of introduced components positively influence intensity of clot formation and its organoleptic characteristics.
The processing of the results of the input selection experiments was based on mathematical statistics methods using a statistical analysis and data visualisation system. This system provides an extensive range of libraries, with functions for comprehensive statistical data analysis, including calculation of the main characteristics of a random sample, such as mathematical expectation, variance, functions of variance analysis, multivariate factor analysis, construction of regression models and many others used in this work. The selected statistical analysis system also provides ample opportunities for visualising the data and the results of the analysis (Annex 1).
The regulating factors were the amount of rennet and bacterial starter:
x₁ - rennet g/100 kg; x₂ - bacterial starter, %;
Controllable factors:
y₁ - active acidity, pH;
y₂ is the titratable acidity, °T. The optimisation conditions are as follows:
y ₁- the active acidity shall not be ≥ the declared value; y₂ - the titratable acidity shall not be ≥ the declared value;
Analysis of the raw data was carried out sequentially for each of the parameters. A multivariate analysis of variance was then conducted to identify the factor having the greatest influence on the sample distribution, and regression models were constructed to describe the dependence of the active acidity of the milk mixture on the parameters studied. Using the regression analysis method, the dependence of active acidity on the amount of starter and the dose of rennet was established. Preliminary correlation matrix was built, complex data obtained in the process of experimental research (Appendix 2). All indicators for each factor were normalised to determine the target function.
The adequacy of the mathematical models was assessed by the value of the coefficient of determination. The square root of the value of the coefficient of determination determines the correlation coefficient. The closer the value of the coefficient of determination is to one, the more adequate the mathematical model is, the smaller the prediction error is.
A surface characterising the target function is plotted for each part of the experiments for the production of soft cheese from a mixture of dairy raw materials. Figures 5, 6 show the dependence of the value of the target function on changes in the amount of sourdough and rennet during acid-soy curdling.

Figure 5 - Variation of active acidity in soft rennet cheeses with changes in the amount of starter and rennet

The regression dependence of this model is as follows:

2
To change the amount of bacterial input закваски ₁z=7,194248+0,₁244445х+0,₂254186х-0,₁²00034х+0,166667x ²+0,₁010604xx₂

1
Дляизмененияколичествавносимогосычужногофермента ₂z=1,835371+0,₁177523х+3,₂954338x-0,0000094х ²-0,₂²25x+0,₁000989xx₂,
The coefficient of determination of the model is 0,971.
The following are slices of the response surfaces (Figure 6) described by the regression equations above.

Figure 6 - Change in active acidity in soft rennet cheeses with changes in the amount of starter and rennet

They show that when the amount of enzyme used is changed, there is a decrease in the active acidity of the solution at any time point. A noticeable outlier in dispersion is observed at the time point of 60 minutes. For a change in the concentration of milk in the solution, the opposite trend is found. As the concentration of goat's milk increases, the active acidity of the solution increases. At the same time, the highest dispersion is observed at time minutes80, regardless of the concentration of goat milk.
A surface characterising the target function is plotted for each part of the experiments for the production of soft cheese from a mixture of dairy raw materials. Figures 7, 8 show the dependence of the value of the target function on the variation of the amount of yeast and rennet during rennet curdling.
The regression dependence of this model is as follows:

1
For starter change z=₃ 5.416667+0.365833x+1₁.5x-0₂.00063x² - 0.33333x+0₂².052x₁₂

2
₄Дляизменениясычужногоферментаz= 3,58333+0,₁200833х+1,₂25х+0,₁²000417х+0,00000621x ²+0,₁0075xx₂,
The coefficient of determination of the model is 0,977.

Figure 7 - Variation of active acidity in soft rennet cheeses with changes in the amount of starter and rennet

Figure 8 - Variation of active acidity in soft rennet cheeses with changes in the amount of starter and rennet

As can be seen from the above dependencies, the maximum value of the target functions is achieved in both cases with the minimum concentration of goat's milk in the mixture and the maximum level of the input component (enzyme and starter).
It was observed that using different doses of rennet was accompanied by whey separation over a wide range of solids content. The intensity of syneresis was influenced by both factors studied (rennet and yeast dosage). An increase in syneresis was observed with an increase in the dose of the bacterial starter.
In analysing the data obtained it was found that the maximum value on the response surfaces corresponds to a dose of 2.5% bacterial starter, 5 g/100 kg rennet and a mixed composition of cow's milk/goat's milk - 50/50.
The data obtained from the studies show that when the amount of rennet was increased, the quality of the clots deteriorated and there was severe dehydration, which led to negative organoleptic properties of the product. Increasing the amount of bacterial yeast resulted in a rapid increase in acidity and subsequently an excessively sour taste to the product.
It can be concluded that the presented mathematical models show the extent to which the quantity of enzyme preparation and bacterial leaven affect the quality of a milk-protein clot, and subsequently the quality of the product.
Based on the presented mathematical analysis of the experimental studies, the optimum dose of rennet in the amount of 5 g/100 kg and bacterial starter - 2.5%, which ensures good organoleptic characteristics of the product and minimal losses of milk components into the whey, was selected for further research.

5.