3.1 Establishment results of standard curve
In order to obtain the standard curve equation for the detection of E. coli in coconut water by Rapid Microbial analyzer, 106-10-1 CFU/mL E. coli bacterial solutions were selected to conduct the experiment. The standard curve was defined and plotted with the logarithm of bacterial count (CFU/mL) -lg CFU/mL as function of the time (h) using the collected data from above tests,, thus the standard curve equation lg C = -0.8460 t+8.6711 (lg C is lg CFU/mL) was obtained with standard deviation (R2) of 0.9856(Figure 3A). R2 is close to 1, indicating that the linear fitting of the equation is desirable, and it can be used as the standard curve equation. Due to Rapid Microbial Analyzer was composed of 16 channels, and the consistency of fluorescence of each channel was tested and compared. As shown from Figure 3B that an excellent consistency among different channels was obtained. Same among analyzers, as illustrated in Figure 3C, although slightly different exist, all the analyzers were within the acceptable range.
3.2 Evaluation Results of Rapid Microbial Analyzer
To verify the standard curve and calibration equation for the Rapid Microbial Analyzer for the detection of E. coli in coconut water, we performed two sets of gradient dilutions of E. coli with known concentrations using negative samples, where the first set of gradient dilutions: The E. coli solution with known concentration was diluted to 0.5×103, 102, 0.5×102, 101, 0.5×101, 100 CFU/mL with negative coconut water; Second group of gradient dilutions: The known concentration E. coli was diluted to 101, 102, 103, 104, 105, 106, 107, 108 CFU/mL with negative coconut water. As comparison, two groups of experiments were conducted according to the national standard [30,31] to evaluate the method. In addition, two instruments were used in both groups to evaluate the consistency and reproducibility of this method.
Table 1. Acceptance results of two groups obtained using national standard method and Rapid Microbial Analyzer
parameters
|
Acceptance standard
|
Group 1 Acceptance Results
|
Group 1 Conclusions
|
Group 2 Acceptance Results
|
Group 2 Conclusions
|
accuracy
|
≥ 95%
|
100%
|
qualified
|
100%
|
qualified
|
sensitivity
|
≥ 95%
|
100%
|
qualified
|
100%
|
qualified
|
specificity
|
≥ 98%
|
100%
|
qualified
|
100%
|
qualified
|
false negative
|
< 5%
|
0%
|
qualified
|
0%
|
qualified
|
false positive
|
< 0%
|
0%
|
qualified
|
0%
|
qualified
|
The results obtained using national standard method for gradient dilution and Rapid Microbial Analyzer were shown in Table1(the acceptance criteria are referring to GB/T 6379, ISO16140 and AOAC), demonstrated that both methods meet the acceptance standard .Meanwhile, we conducted linear comparison on the bacterial counts measured by two methods. As addressed in Figure 4A and Figure 4B, the specific linear relationship showed that both of the correlation coefficients were 0.999. Once again validated Rapid Microbial Analyzer measurements meet the national standard. The instrument consistency measurement with different Rapid Microbial Analyzers also performed, as indicated in Figure 4C and 4D, a desired result was achieved, which was consistent with the conclusion we obtained from Figure 3C. Same with the good reproducibility result addressed in Figure 4E and 4F. Thus, we can conclude that the Rapid Microbial Analyzers show good consistency and reproducibility at low and high concentrations, andare suitable for the detection of E. coli in coconut water.
3.3 Performance verification of Rapid Microbial Analyzer for actual samples
To study the specificity of Rapid Microbial Analyzer for Escherichia coli measurement, three more bacteria (Staphylococcus aureus, Salmonella typhi and Pseudomonas aeruginosa) were selected as comparison , As demonstrated in Figure 5A, the detection for Escherichia coli showed a high and strong fluorescence signal, while barely observed signal for other three miscellaneous bacteria, suggesting an excellent specificity of Rapid Microbial Analyzer for sensing Escherichia coli in coconut water. To further challenge the specificity of Rapid Microbial Analyzer, E.coli was mixed with S.aureus and S.typhi. As shown in Figure 5C, no interference can be observed from other bacteria, which once again suggesting an excellent specificity of Rapid Microbial Analyzer. The sensitivity test results of Escherichia coli in coconut water detected by Rapid Microbial Analyzer are shown in Figure 5B, and it shows that the limit-of-detection of 1 CFU /mL can be achieved, indicating a high sensitivity of the instrument.
Finally, we challenged our Rapid Microbial Analyzer with real coconut water samples obtained from different sources for sensing Escherichia coli. As shown in Figure 5D, Sample 2 was positive, while Sample 1 and Sample 3 were negative. Meanwhile, the number of Escherichia coli colonies obtained by plate counting from the diluted actual sample (Table S1) was 106 CFU/mL, which was in the same order of magnitude with that obtained from Rapid Microbial Analyzer method, and it could be concluded that the Rapid Microbial Analyzer can be used for practical detection of Escherichia coli in coconut water.