Comparison among the Quantification of Bacterial Pathogens by qPCR, dPCR, and Cultural Methods

Abstract

The demand for rapid methods for the quantification of pathogens is increasing. Among these methods, those based on nucleic acids amplification (quantitative PCRs) are the most widespread worldwide. Together with the qPCR, a new approach named digital PCR (dPCR), has rapidly gained importance. The aim of our study was to compare the results obtained using two different dPCR systems and one qPCR in the quantification of three different bacterial pathogens: Listeria monocytogenes, Francisella tularensis, and Mycobacterium avium subsp. paratuberculosis. For this purpose, three pre-existing qPCRs were used, while the same primers and probes, as well as PCR conditions, were transferred to two different dPCR systems: the QX200 (Bio-Rad) and the Quant Studio 3D (Applied Biosystems). The limits of detection and limits of quantification for all pathogens, and all PCR approaches applied, were determined using genomic pure DNAs. The quantification of unknown decimal suspensions of the three bacteria obtained by the three different PCR approaches was compared through the Linear Regression and Bland and Altman analyses. Our results suggest that, both dPCRs are able to quantify the same amount of bacteria, while the comparison among dPCRs and qPCRs, showed both over and under-estimation of the bacteria present in the unknown suspensions. Our results showed qPCR over-estimated the amount of M. avium subsp. paratuberculosis and F. tularensis cells. On the contrary, qPCR, compared to QX200 dPCR, under-estimated the amount of L. monocytogenes cells. However, the maximum difference among PCRs approaches was <0.5 Log₁₀, while cultural methods underestimated the number of bacteria by one to two Log₁₀ for Francisella tularensis and Mycobacterium avium subsp. paratuberculosis. On the other hand, cultural and PCRs methods quantified the same amount of bacteria for L. monocytogenes, suggesting for this last pathogen, PCRs approaches can be considered as a valid alternative to the cultural ones.

Keywords: bacteria; dPCR; pathogens; qPCR; quantification.

Figure 1

Linear regressions and Logit analysis plots of the three qPCRs used in the study. Panels (A,D) are referred to Mycobacterium avium subsp. paratuberculosis (MAP), panels (B,E) are referred to Listeria monocytogenes (Lm) and panels (C,F) to Francisella tularensis (Ft). The number of replicates used for the evaluation of the panels (A–C) is reported in material and method section. In the right side of each graph are reported the first-grade equation as well as the r² values. ^*Indicates 2/10 replicates were positive. Bars represent standard deviation. Panels (D–F) show the LOD_95% for each qPCR, which are the minimum amounts of DNA detectable with a 95% probability.

Figure 2

Linear regressions and Bland and Altman analyses of Mycobacterium avium subsp. paratuberculosis (MAP) cells in unknown samples obtained by dPCRs (QuantStudio 3D and QX200) and qPCR. (A) Linear regression between the two dPCRs; (B) Linear regression of dPCR QuantStudio 3D and qPCR; (C) Bland and Altman analysis of dPCR QuantStudio 3D and qPCR; (D) Linear regression of dPCR QX200 and qPCR; (E) Bland and Altman analysis of dPCR QX200 and qPCR. Bland and Altman analysis between the two dPCRs platform was not possible because the differences were not normally distributed. Data were reported as Log₁₀ cells per μL. The gray lines in linear regression plots represent the ideal regression value.

Figure 3

Linear regressions and Bland and Altman analyses of L. monocytogenes (Lm) cells in unknown samples obtained by dPCRs (QuantStudio 3D and QX200) and qPCR. (A) Linear regression between the two dPCRs; (B) Bland and Altman analysis between the two dPCRs; (C) Linear regression of dPCR QuantStudio 3D and qPCR; (D) Linear regression of dPCR QX200 and qPCR; (E) Bland and Altman analysis of dPCR QX200 and qPCR. Bland and Altman analysis between QuantStudio 3D and qPCR was not possible because the differences were not normally distributed. Data were reported as Log₁₀ cells per μL. The gray lines in linear regression plots represent the ideal regression value.

Figure 4

Linear regressions and Bland and Altman analyses of F. tularensis (Ft) cells in unknown samples obtained by dPCRs (QuantStudio 3D and QX200) and qPCR. (A) Linear regression between the two dPCRs; (B) Bland and Altman analysis between the two dPCRs; (C) Linear regression of dPCR QuantStudio 3D and qPCR; (D) Bland and Altman analysis of dPCR QuantStudio 3D and qPCR; (E) Linear regression of dPCR QX200 and qPCR; (F) Bland and Altman analysis of dPCR QX200 and qPCR. Data were reported as Log₁₀ cells per μL. The gray lines in linear regression plots represent the ideal regression value.