D-Count Device Qualification Protocol Provided by BioMérieux
生物梅里埃提供的 D-Count 设备认证方案

1. Objective
1. 目标

To compare the performance of the flow cytometry method (D-Count) with the reference method (petri dish count) for the quantification of microorganisms in dairy and similar products.
比较流式细胞术法 （D-Count） 与参比法 （培养皿计数） 在定量乳制品和类似产品中微生物方面的性能。

Reference Method: NF EN ISO 16140-2 - September 2016. "Microbiology of the Food Chain - Method Validation - Part 2: Protocol for the Validation of Alternative (Commercial) Methods Against a Reference Method"
参考方法：NF EN ISO 16140-2 - 2016 年 9 月。“食物链的微生物学 - 方法验证 - 第 2 部分：根据参考方法验证替代（商业）方法的方案”

2. Product Categories and Number of Samples
2. 产品类别和样品数量

Neutral pH (4 types): Gourmet Island, Whipping Cream, Yop Chocolate, Sugar Syrup
中性 pH 值（4 种）：Gourmet Island、淡奶油、Yop 巧克力、糖浆

(5 samples per type, 3 contamination levels, duplicate measurements = 60 samples * 2 measurements = 120 analyses per method).
（每种类型 5 个样品，3 个污染水平，重复测量 = 60 个样品 * 2 次测量 = 每种方法 120 次分析 ）。

Acidic pH (6 types): Skyr Fruit Bed, Milk Pearl mixed with Fruit Puree, Cream Cheese, Extra Cuddle, Fruit Yogurt, Drinking Yogurt.
酸性 pH 值（6 种）：Skyr Fruit Bed、牛奶珍珠与水果泥混合、奶油奶酪、Extra Cuddle、水果酸奶、饮用酸奶。

(5 samples per type, 3 contamination levels, duplicate measurements = 90 samples * 2 measurements = 180 analyses per method)
（每种类型 5 个样品，3 个污染水平，重复测量 = 90 个样品 * 2 次测量 = 每种方法 180 次分析 ）

Total analyses by product category
按产品类别划分的分析总数

240 analyses (cytometry and spread on a dish) for acidic pH
240 次酸性 pH 值的分析（细胞分析和铺展在培养皿上）

360 analyses for acidic pH
酸性 pH 值的 360 度分析

Total : 600 analyses.
总计：600 次分析。

For each type, it is not necessary for the products to come from the same production. On the contrary, it will be more interesting to carry out the study on different types (e.g. different tastes) as long as the traceability of the exact product used is available.
对于每种类型，产品不必来自同一生产。相反，只要可以使用确切产品的可追溯性，对不同类型（例如不同口味）进行研究会更有趣。

3. Contamination Levels
3. 污染水平

Three levels of contamination will be tested:
将测试三个级别的污染：

0 CFU/pot
0 CFU/底池

1 - 10 CFU/pot
1 - 10 CFU/锅

10 – 100 CFU/pot
10 – 100 CFU/罐

For each level of contamination: the concentration must be checked with double spread on a petri dish.
对于每个污染水平：必须在培养皿上双铺展检查浓度。

4. Sample Preparation
4. 样品制备

Microorganisms inoculated:
接种的微生物：

Yeasts and Moulds: Strains already isolated on site and identified.
酵母菌和霉菌： 已在现场分离并鉴定的菌株。

Make a mixture of the 3 stumps. The final concentration in the product is the concentration of the 3 strains combined.
将 3 个树桩混合。 产品中的最终浓度是 3 种菌株组合的浓度。

FMAR (Revivable Aerobic Mesophilic Flora): Recovery from a raw milk sample according to the following protocol:
FMAR（可存活的需氧嗜温菌群）： 根据以下方案从生牛奶样品中回收：

Cascading dilution of raw milk.
生牛奶的级联稀释。

Spread each dilution on a Petri dish.
将每种稀释液涂抹在培养皿上。

Selecting a countable box.
选择可计数的框。

Calculation of the CFU/mL concentration of raw milk taking into account the dilution factor.
考虑稀释因子计算原料奶的 CFU/mL 浓度。

Identification of germs is not necessary.
无需鉴定病菌。

Inoculation: Samples will be inoculated with the mixture of yeast/mold strains and MARL at concentrations corresponding to the three levels of contamination.
接种： 样品将接种酵母/霉菌菌株和 MARL 的混合物，其浓度对应于三个污染水平。

Pre-incubation:
预孵育：

On the entire product
在整个产品上

Yeasts / moulds: 24 hours at a temperature of 25 °C.
酵母菌/霉菌： 在 25 °C 的温度下放置 24 小时。

MARL: 24 hours at a temperature of 30°C.
MARL： 在 30°C 的温度下放置 24 小时。

5. Enumeration
5. 枚举

Reference Method: Petri dish count (CFU/g) after incubation.
参考方法： 孵育后培养皿计数 （CFU/g）。

YGC for yeast/mold
用于酵母/霉菌的 YGC

BCP for MARLANT
MARLANT 的 BCP

D-Count method: Count of the number of viable cells (CFU/g) by incubation flow cytometry.
D-Count 方法： 通过孵育流式细胞术对活细胞数 （CFU/g） 进行计数。

6. Processing of Results
6. 结果处理

Quantitative Analysis: The results will be statistically analyzed to determine the standard deviation of qualification and the accuracy of the response of the D-Count method compared to the reference method. The statistical analysis will include:
定量分析： 将对结果进行统计分析，以确定与参考方法相比，D-Count 方法的定性标准差和响应的准确性。统计分析将包括：

Averaging the differences between the two methods
平均两种方法之间的差异

Calculating the standard deviation of the two methods
计算两种方法的标准差

For every level of contamination and every product.
适用于各种污染级别和每种产品。

Student's t-test to compare the means of the two methods.
用于比较两种方法的均值的学生 t 检验。

Qualitative Analysis: If the results do not lend themselves to quantitative analysis, a qualitative approach will be used to assess the agreement between the two methods using "positive/negative" categories.
定性分析： 如果结果不适合定量分析，将使用定性方法使用 “积极/消极 ”类别来评估两种方法之间的一致性。

7. Qualification Criteria
7. 资格标准

The D-Count method will be considered qualified if it meets the following criteria:
如果 D-Count 方法满足以下标准，则认为该方法合格：

Accuracy: The bias between the two methods must be below a predefined threshold, set at a maximum of 1 log.
准确性： 两种方法之间的偏差必须低于预定义的阈值，最大值设置为 1 个对数。

Accuracy: The repeatability of the D-Count method must be below predefined thresholds, set at 0.25.
准确度：D-Count 方法的重复性必须低于预定义的阈值，设置为 0.25。

8. Results
8. 结果

The table below summarizes all the results of the analyses carried out on unseeded products.
下表总结了对非种子产品执行的所有分析结果。

Microorganisms are not detected on any analysis, either in the traditional method or in the cyto method.
无论是在传统方法还是细胞法中，任何分析都无法检测到微生物。

In the following table, we have grouped the data set by matrix and level of contamination.
在下表中，我们按基质和污染水平对数据集进行了分组。

There is an overall underestimation of the D count method of around 0.67 log compared to the box method, without this being critical, knowing that both methods detect 100% of contaminations.
与盒法相比，D 计数法总体上低估了大约 0.67 个对数，这并不重要，因为两种方法都检测到 100% 的污染。

In the following table, we have grouped this data by matrix:
在下表中，我们按矩阵对这些数据进行了分组：

We can conclude from these different data that regardless of the matrix or the concentration, the per cyto method slightly underestimates the final rate but within the limit <1log and with a sensitivity equivalent to the box method.
从这些不同的数据中我们可以得出结论，无论基质或浓度如何，每细胞计数方法都略微低估了最终速率，但在 <1log 的限值范围内，并且具有与盒法相当的灵敏度。

Finally, in the table below, we have grouped the standard deviations of the repeatability of the methods, by matrix and by concentration:
最后，在下表中，我们按基质和浓度对方法重复性的标准差进行了分组：

These results show a mean standard deviation of 0.15 larger for the D-Count method, which remains within our acceptability tolerance (<0.25).
这些结果表明，D-Count 方法的平均标准差高出 0.15，这仍然在我们的可接受容差范围内 （<0.25）。

In parallel with these analyses, in order to determine the repeatability of the D-Count method, we performed 5 batches with 23 tubes made from the same sample.
在进行这些分析的同时，为了确定 D-Count 方法的可重复性，我们进行了 5 个批次，使用由同一样品制成的 23 支试管。

We used different contamination rates to see if the repeatability is the same depending on the load being analyzed. In view of the above results, we can conclude that regardless of the initial load, the D-count method has an equivalent repeatability and an uncertainty of 0.22 log (2EC).
我们使用不同的污染率来查看重复性是否相同，具体取决于所分析的载荷。鉴于上述结果，我们可以得出结论，无论初始载荷如何，D-count 方法都具有等效的重复性和 0.22 log （2EC） 的不确定性。