Real-time PCR based quantitative microbial source tracking : method characteristics and applicability

Abstract

Recent years have brought huge advances in methods for the identification of the sources of faecal pollution. However the use of quantitative real-time PCR (qPCR) for quantitative microbial source tracking (QMST) is facing challenges on the levels of methodology and applied study design. In terms of methodology this study aimed at defining some key issues for the quality assurance of QMST results. These include necessary definitions of the limit of detection (LOD) and the limit of quantification (LOQ) in qPCR methods based on the characteristics of the acquired data. Important factors influencing these key figures are the stochastic and truncated distribution of qPCR data at low concentrations and the heteroscedasticity of the measurement variability. In the light of the recovered results, further investigations should concentrate on the development of flexible confidence intervals, which take negative results into consideration or a probabilistic approach to determine the probability of a correct measurement, given a certain amount of target.<br />In addition DNA extraction efficiencies were determined for QMST markers in various environmental water samples which help in determining sample limits of detection (SLODs). To assess the applicability of the ruminant-specific QMST BacR and the human-specific BacH methods beyond the catchment area they were originally developed in LKAS2, BacR/H were tested on water samples from alpine karstic spring LKAS8 situated on the Eastern verges of the Austrian calcareous alps. During a period of 17 months QMST markers, classical faecal indicators, hydrological and chemo-physical parameters were investigated. Taking into account the hydrological and water quality characteristics of the spring an innovative nested sampling design with three tiers was applied: (i) a three-weekly basic monitoring, (ii) a high frequency monitoring with 2 samplings per week during summer months, and (iii) an investigation of a late summer flood event. Statistical analysis of the results revealed the strong role of ruminant sources in faecal pollution of the spring while human sources seem to have a negligible impact on the spring water quality. This was especially evident in samples taken during the summer months and the flood event. In addition the nested sampling design proved to be a faesible and comprehensive approach for QMST investigations.The results from this study are in strong accordance with those from a previous study on the karstic spring LKAS2, emphasising the usefulness of the applied QMST approach.