Skip to content
Resources > Latest News > Bacteroidales revolutionising microbial source tracking

Bacteroidales: Revolutionising microbial source tracking

By B J Tillett and V Pettigrove.

First published in Water e-Journal Vol 2 No 1 2017.



The age of using Escherichia coli and Enterococci sp. as standard faecal indicator bacteria is coming to end as new technological advancements enable us to identify which animals (i.e. human, birds, dogs, cattle etc.) are contributing faecal contamination in waterways and therefore provide a better understanding of the risks posed by this type of pollution.

Bacteroidales are increasingly being used to source human faecal pollution in urban streams and stormwater drains. Here, we clearly outline their favourable characteristics and provide a detailed discussion of analytical methods. Examples of their successful application in waterway assessments are also provided.


Growing populations in urbanised, coastal areas are increasing pressure on water managers to provide safe recreational waters (Image 1). Escherichia coli and Enterococci sp. have long been preferred faecal indicators in water quality assessments because of their high abundance in the faeces of warm blooded mammals (McLellan & Eren, 2014).

Until the last few decades, it was widely accepted that quantification of E. coli was a satisfactory indicator of human health risk. Now, with advances in pathology, it is becoming evident that different sources of faecal contamination (i.e. human and animals), vary in the occurrence of zoonoses and consequently risk to human health. For example, dog faeces can transmit Toxocariasis infections in humans. This parasite can cause serious illness and blindness (Robertson & Thompson, 2002).

Similarly, cow faeces can transmit Leptospirosis in humans which can cause severe liver and kidney disease (Azizi et al., 2012). Correct sourcing of faecal contamination can not only inform on whether faecal pollution is present, but also likely health risk associated with contamination. Unfortunately, Escherichia coli and Enterococci sp. lack the host specificity required to discriminate human and animal faecal sources.

Another issue with standard faecal indicator bacteria is increasing evidence of environmental Escherichia coli and Enterococci sp. strains that have no faecal origin and which cannot be discriminated by current culture methods, reducing confidence that recorded levels are truly representative of faecal contamination (Power et al., 2005, Byappanahalli et al., 2012, Lleo et al., 2005).

Many scientists and managers alike aver that current bacterial assessments are inappropriate, given the challenges and complexity of pollution sourcing; calling for more informative and rapid indicators of faecal contamination.

Bacteroidales: A faecal indicator of the future

New indicator bacteria that are increasing in popularity belong to the Archaean order Bacteroidales (Dick & Field, 2004, Stea et al., 2015, Walker et al., 2015). Bacteroidales, unlike traditional indicator bacteria are obligate anaerobes and therefore are unlikely to propagate outside an animal’s digestive system (Scott et al., 2002).

Image 1. Beachgoers enjoying a warm summers day on a beach in Melbourne, Victoria

They are more abundant than either Escherichia coli or Enterococci sp. in faeces. Arguably, their greatest attribute is that they display high specificity, coevolving within their hosts’ digestive system (Dick et al., 2005, Witty et al., 2009).

These assays reflect a change in protocol away from culture dependent assessments towards the application of genetic technologies, specifically polymerase chain reaction methods (PCR), to amplify bacteria to a quantifiable level i.e. quantitative-PCR or real-time PCR (qPCR and rt-PCR respectively). These methodologies have the potential to revolutionise microbial source tracking in waterways because the choice of indicator bacteria is no longer limited to those which can be easily cultured, but rather select bacteria that have desired traits such as high host-specificity and abundance (Roslev & Bukh, 2011, Witty et al., 2009).

Another major advantage of genetic amplification methods is greater quantification of amplification efficiency. In culture-based approaches, it is assumed that bacterial levels successfully grown on culture medium directly correlate to environmental levels. Often, limited quality control measures are reported. Conversely, qPCR assays are specifically designed for each species marker. Each assays comprises a unique set of reagents and primers which are subjected to a pre-defined series of thermo-cycles i.e. hot / cold cycles to maximise amplification of target bacteria (Bernhard & Field, 2000).

During this process, target bacteria are fluorescently tagged and their associated fluorescence quantified in real-time during each thermo-cycle. The qPCR light cycler machine quantifies “crossing points” (Cp) values which are defined by the manufacturer as the cycle at which the fluorescence of a sample exceeds background levels i.e. the number of light cycles until the sample concentration is reliably detected. These Cp values are then compared against a standard curve and converted to concentrations (Dick & Field, 2004).

The standard curve is generated from replicate serial dilutions of known-concentration reference standards. Here, light cycler-generated Cp values are plotted against predetermined concentrations and a linear regression fit using the second derivative method.

The range of the standard curve should encompass the expected distribution of the samples to be tested to ensure accurate quantification. The fit of the data to the standard curve is then used to generate quality control indices. Broad assay performance indices report the accuracy of lightcycler quantification by assessing the deviation of light cycler-determined concentrations of standards from predefined concentrations and also amplification efficiency.

If the reaction is 100% efficient, we would expect the number of fluorescently tagged bacteria to exactly double each cycle representing a value of “2”.

Generation of a standard curve for each speciesmarker also allows for distinct definitions of detection and quantification limits which provide the user with a complete understanding of the power and error within each species marker.

The limit of detection is commonly determined from the linear equation as the concentration when the Cp value is zero or a pre-set threshold i.e. five cycles. Another approach is to analyse serial dilutions until the light cycler can no longer detect any target DNA beyond background levels. The limit of quantification can be determined as the exact concentration at which the linear relationship between concentration and Cp value breaks down.

Another common approach uses ANOVA to determine the concentration which can no longer reliably be discriminated from previous levels.

Additional control measures are also added to assays. These commonly include: analysing replicate samples where the average Cp or concentration and standard deviations are reported; negative filtration controls (filtration of distilled water and extraction of DNA from associated membrane); negative PCR controls (replacement of environmental water sample with distilled water in certain PCR reactions); positive PCR controls (amplification of multiple known-concentration reference standards); internal amplification controls (all reactions are spiked with a reference material to determine if PCR inhibitors are present within each reaction); and recovery efficiency indicators/extraction control (each environmental water sample is spiked with a known concentration of non-target DNA, typically salmon sperm DNA, during the extraction process which is quantified in a separate qPCR assay).

Comprehensive research has investigated marker specificity (amplification in non-target species) and sensitivity (amplification in target species) to verify that the most accurate and reliable segment of Bacteroides sp. 16s rRNA or other gut fauna in the case of birds which have low levels of Bacteroides sp., have been targeted (Ahmed et al., 2012, Dick et al., 2005, Fogarty & Voytek, 2005).

Results from one of these studies are provided in Table 1 (Kildare et al., 2007). If a marker is shown to cross-amplify in a non-target species, a major advantage of qPCR assays is that the limit of quantification for the target species can be altered accordingly to exclude these incorrect positive results.

Bayes’ Theorem can also be applied combining assay sensitivity and specificity, with background levels (determined as the frequency of environmental samples tested with target concentrations above the limit of quantification) to determine the likelihood of results representing true positives.

Distribution and abundance of markers within host populations across different geographic ranges and life history stages have also been tested (Ahmed et al., 2008, Okabe et al., 2007, Shanks et al., 2014). Region-specific variation in marker sensitivity and specificity confirm it is paramount that the distribution and abundance of Bacteroides sp. markers in local populations be determined prior to any investigations.

At The University of Melbourne’s Centre for Aquatic Pollution and Identification Management (CAPIM), we have tested the presence and cross amplification of human, dog, seagull and ruminant markers in a small subsample from Victoria (Table 2).

Our results support other research in the literature which confirmed the presence of high concentrations of human Bacteroides sp. in wastewater and stormwater across Australia (Ahmed et al., 2016, Sidhu et al., 2013), reiterating their applicability and value in Australian local water quality assessments. Age-related differences in Bacteroides sp. present in host populations are less clear but differences are suggested to be most prevalent in species with long weaning time for young, such as cattle (Shanks et al., 2014). Long-term agricultural feeding strategies have also been shown to influence marker sensitivity in cattle (Shanks et al., 2011). However, the presence of specific Bacteroides sp. is not only based on what food is digested, but also how food is digested. The strength of these qPCR assays is that these sources of error / variation can be quantified in pilot studies and the parameters of the assays adjusted accordingly to increase overall accuracy and confidence of results.

A standard protocol is currently being developed by the US EPA for sourcing human faecal contamination. Unique Bacteroides sp. have also been identified from ruminants, canine, seagull and duck digestive systems among others, although these have not yet been adapted into standard protocols.

Arguably one of the main challenges for the use of Bacteroides sp. markers to date is ensuring repeatability and reproducibility of results. Shanks et al., (2016) have outlined data acceptance criteria for standardised human-associated faecal source identification based on a multi-lab investigation.

Modern sourcing of microbial pollution in waterways requires a complex suite of indicator organisms to equal the ever increasing number of mixed point and non-point sources of pollution. Bacteroides sp. are showing real promise in their successful sourcing of microbial pollution, particularly between human and non-human sources.

Their value is highest in combination with traditional faecal indicator bacteria such as Escherichia coli or Enterococci sp. and chemical markers in a multi-tiered approach to microbial source tracking.

Successful applications of Bacteroides to source human and on-human faecal contribution

Staff from CAPIM in collaboration with local councils, Melbourne Water and EPA Victoria have successfully used Bacteroides sp. assays in studies to determine the presence of human faecal contamination in rivers and estuaries.

Bacteroides sp. assays in conjunction with traditional faecal indicator bacteria Escherichia coli and the chemical indicator ammonia have been used to source human faecal pollution in urban waterways within Melbourne city, including Dandenong Creek and Merri Creek and Frankston and Mornington shires on the Mornington Peninsular.

These assays have also been undertaken in the rural catchments of Wye River, The Otways Coast and Werribee, located 32 km south-west of Melbourne, Victoria.

The inclusion of a general Bacteroides sp. marker which detects all sources of faecal contamination, not just those of concern confirmed that faecal contamination was present in Dandenong Creek.

The subsequent use of species-specific Bacteroides sp. markers (dog, human and ruminant) showed that the major source of this contamination was non-human (dog and ruminant) although strong human-specific markers were present at some sites.

Measurements using Escherichia coli showed that faecal contamination was above recommended levels for recreational swimming. However Bacteroides sp. markers identified low levels of faecal contamination that was not detected by E. coli assessments (Kellar et al., 2016).

A multi-tiered approach was implemented in the Frankston and Mornington surveys where we were able to use a combination of Escherichia coli, Bacteroides sp. and chemical ammonia indicators to identify human faecal contamination in streams (Figure 1).

These were then traced up the stormwater drains using real-time AMI Ammonia test kits® and Bacteroides sp. testing was used to confirm whether any faecal contamination was of human origin. We were able to identify three separate locations where cracked sewerage infrastructure was leaking untreated sewage directly into stormwater.

Once these leakages had been rectified, human faecal contamination was no longer detected. Such leakages can be identified by other approaches such as directly inserting high-tech cameras into stormwater drains and looking for visual evidence of cracks.

But such approaches are often extremely costly for local councils and often require timely surveys through complex networks before cracks are located.

The advantage of combining Bacteroides sp. assays with a targeted sampling approach is that large urbanised areas can be quickly monitored and key areas flagged for further inspection. On multiple occasions throughout this survey, we were able to source elevated E. coli levels as non-human derived, identifying them as from either seagull or dog faeces, confirming non-point source contamination (Sharley et al., 2014)

Figure 1. Microbial source tracking using multiple indices (chemical, Bacteroidales sp. & Escherichia coli) in Frankston, Victoria. Non-human and human sources were identified. Cracked sewer pipes were found at sites 790, 891 and 892

A second follow-up study was conducted again in Frankston. Similar faecal indicators were employed and again we were able to confirm that elevated E. coli levels were sourced from humans in the lower catchment and were able to identify an emergency spill and another damaged sewer discharging raw sewage into stormwater (Amos & Sharley, 2016).

The application of Bacteroides sp. markers in a rural catchment was able to confirm likely agricultural sources of contamination. In Wye River, sites were tested for human, ruminant, duck and universal Bacteroides sp. markers (Figure 2).

Results from the universal marker again confirmed faecal contamination at multiple sites. Both the duck and human markers identified levels of contamination that were below the limit of quantification indicating that faecal contributions from these sources are low. Conversely, the ruminant marker identified quantifiable levels of faecal contamination and, interestingly, the levels of contamination identified at each site was comparatively similar to those identified by the universal marker, albeit an order of magnitude less, suggesting that ruminant sources are likely to be a major contributor of faecal contamination at each site.

These results suggest a lower risk to human health than if the contamination was sourced as of human origin. In the Werribee surveys, the same markers were tested and again, Bacteroides sp. assays identified low-levels of dog and ruminant faecal contamination not detected by Escherichia coli and confirmed the high levels of E. coli recorded at one site was of human origin, suggesting a potential cracked sewer leaking into stormwater

Figure 2. Wye River Bacteroides sp. assessment results. Average qPCR light cycler derived crossing point values (defined as the number of polymerase chain reaction (PCR) cycles until the sample fluorescence recorded by the PCR machine exceeds background levels) with associated standard deviations are reported. Human and duck species markers were below the limit of quantification i.e. recorded levels are too low to be reliably quantified, and are in grey scale. Both the universal and ruminant markers recorded quantifiable levels of faecal contamination and displayed similar patterns across sampled sites.

The examples described above showcase the application of Bacteroides sp. assays in both snapshot and longer-term assessments of faecal contamination. In snapshot assessments, Bacteroides sp. assays provide a powerful, highly sensitive and specific approach to reliably source faecal contamination at a particular point in time.

In longer-term assessments, when combined with a targeted sampling strategy faecal signals can be traced within catchments to identify point sources and temporal changes in specific faecal sources of contamination, as opposed to general measures as currently provided by standard E. coli counts alone.

Exactly what environmental levels of Bacteroides sp. correlate to human health concerns is not yet defined for all pathogens. However, this approach does allow managers to prioritise risk based on whether the faecal contamination is human or animal and more specifically ruminant, bird or domestic. Each faecal source poses different and known risks to human health based on the presence of specific zoonotic pathogens.


The sole use of general indicators such as E.coli and Enterococci sp. is no longer sufficient to infer human faecal pollution in urban waterways. These standard faecal indicator bacteria are incapable of discriminating between the growing number of human and non-human sources of faecal contamination and consequently differing risks to human health. Furthermore, poor identification of contamination source restricts appropriate management to ensure the contamination event does not reoccur.

Bacteroides sp. can provide information on what animals are the sources of any observed faecal pollution and therefore provide managers with a better understanding of the risk posed by this contamination to human health and best management practices. Sourcing faecal contamination through complex stormwater systems requires an equally diverse toolbox of which Bacteroides sp. are proving to be a much valued, additional source information, than could be provided by E. coli counts alone.

Combined with a target sampling strategy, Bacteroides sp. assays provide a comparatively cheap and quick method to monitor both point and non-point sources in large, often urbanised catchments.

About the authors

Dr Bree Tillett, PhD | Bree is a research fellow at the Centre for Aquatic Pollution Identification and Management (CAPIM) at the University of Melbourne. Her research focuses on the application of genetics technologies to source faecal pollution in waterways.

Vincent Pettigrove | Vincent is the Chief Executive Officer of CAPIM and Principal Research Fellow at The University of Melbourne, Australia. He is a Principal Ecologist at Melbourne Water with over 30 years’ experience in the Victorian water industry and aquatic research.


Ahmed W, Masters N & Toze S (2012): Consistency in the host specificity and host sensitivity of the Bacteroides HF183 marker for sewage pollution tracking. Letters in Applied Microbiology, 55, pp 283-289.
Ahmed W, Sidhu JPS, Smith K, Beale DJ, Gyawali P & Toze S (2016): Distributions of Fecal Markers in Wastewater from Different Climatic Zones for Human Fecal Pollution Tracking in Australian Surface Waters. Applied and Environmental Microbiology, 82, pp 1316-1323.
Ahmed W, Stewart J, Powell D & Gardner T (2008): Evaluation of Bacteroides markers for the detection of human faecal pollution. Letters in Applied Microbiology, 46, pp 237-242.
Amos C & Sharley D (2016): Frankston Microbial Sourcing 2015-2016, Centre for Aquatic Pollution Identification And Management, Technical Report No. 66, University of Melbourne, Victoria, Australia.
Azizi S, Tajbakhsh E, Hajimirzaei MR, Varnamkhasti MG, Sadeghian H & Oryan A (2012): Evaluation of ‘white-spotted kidneys’ associated with leptospirosis by polymerase chain reaction based LipL32 gene in slaughtered cows. Journal of the South African Veterinary Association-Tydskrif Van Die Suid-Afrikaanse Veterinere Vereniging, 83.
Bernhard AE & Field KG (2000): A PCR assay to discriminate human and ruminant feces on the basis of host differences in Bacteroides-Prevotella genes encoding 16S rRNA. Applied and Environmental Microbiology, 66, pp 4571-4574.
Byappanahalli MN, Nevers MB, Korajkic A, Staley ZR & Harwood VJ (2012): Enterococci in the Environment. Microbiology and Molecular Biology Reviews, 76, pp 685-706.
Dick LK, Bernhard AE, Brodeur TJ, Domingo JWS, Simpson JM, Walters SP & Field KG (2005): Host distributions of uncultivated fecal Bacteroidales bacteria reveal genetic markers for fecal source identification. Applied and Environmental Microbiology, 71, pp 3184-3191.
Dick LK & Field KG (2004): Rapid estimation of numbers of fecal Bacteroidetes by use of a quantitative PCR assay for 16S rNA genes. Applied and Environmental Microbiology, 70, pp 5695-5697.
Fogarty LR & Voytek MA (2005): Comparison of BacteroidesPrevotella 16S rRNA genetic markers for fecal samples from different animal species. Applied and Environmental Microbiology, 71, pp 5999-6007.
Kellar C, Sharley D, Tillett B, Marshall S, Long S & Pettigrove V (2016): Enhancing Our Dandenong Creek Summary Report for 2015 – 2016, Centre for Aquatic Pollution Identification And Management, Technical Report No.68, University of Melbourne, Victoria, Australia.
Kildare BJ, Leutenegger CM, Mcswain BS, Bambic DG, Rajal VB & Wuertz S (2007): 16S rRNA-based assays for quantitative detection of universal, human-, cow-, and dogspecific fecal Bacteroidales: A Bayesian approach. Water Research, 41, pp 3701-3715.
Lleo MD, Bonato B, Benedetti D & Canepari P (2005): Survival of enterococcal species in aquatic environments. Fems Microbiology Ecology, 54, pp 189-196.
Mclellan SL & Eren AM (2014): Discovering new indicators of fecal pollution. Trends in Microbiology, 22, pp 697-706.
Okabe S, Okayama N, Savichtcheva O & Ito T (2007): Quantification of host-specific Bacteroides-Prevotella 16S rRNA genetic markers for assessment of fecal pollution in freshwater. Applied Microbiology and Biotechnology, 74, pp 890-901.
Power ML, Littlefield-Wyer J, Gordon DM, Veal DA & Slade MB (2005): Phenotypic and genotypic characterization of encapsulated Escherichia coli isolated from blooms in two Australian lakes. Environmental Microbiology, 7, pp 631-640.
Robertson ID & Thompson RC (2002): Enteric parasitic zoonoses of domesticated dogs and cats. Microbes and Infection, 4, pp 867-873.
Roslev P & Bukh AS (2011): State of the art molecular markers for fecal pollution source tracking in water. Applied Microbiology and Biotechnology, 89, pp 1341-1355.
Scott TM, Rose JB, Jenkins TM, Farrah SR & Lukasik J (2002): Microbial source tracking: Current methodology and future directions. Applied and Environmental Microbiology, 68, pp 5796-5803.
Shanks OC, Kelty CA, Archibeque S, Jenkins M, Newton RJ, Mclellan SL, Huse SM & Sogin ML (2011): Community Structures of Fecal Bacteria in Cattle from Different Animal Feeding Operations. Applied and Environmental Microbiology, 77, pp 2992-3001.
Shanks OC, Kelty CA, Peed L, Sivaganesan M, Mooney T & Jenkins M (2014): Age-Related Shifts in the Density and Distribution of Genetic Marker Water Quality Indicators in Cow and Calf Feces. Applied and Environmental Microbiology, 80, pp 1588-1594.
Sharley D, Valenzuela I, Sharp S, Amos C, Sa Almeida I & Pettigrove V (2014): Identifying the origin of microbial contamination at Frankston and Mornington beaches, Centre for Aquatic Pollution Identification And Management, Technical Report No. 35, University of Melbourne, Victoria, Australia.
Sidhu JPS, Ahmed W, Gernjak W, Aryal R, Mccarthy D, Palmer A, Kolotelo P & Toze S (2013): Sewage pollution in urban stormwater runoff as evident from the widespread presence of multiple microbial and chemical source tracking markers. Science of the Total Environment, 463, pp 488-496.
Stea EC, Hansen LT, Jamieson RC & Yost CK (2015): Fecal Contamination in the Surface Waters of a Rural- and an Urban-Source Watershed. Journal of Environmental Quality, 44, pp 1556-1567.
Walker JW, Van Duivenboden R & Neale MW (2015): A tiered approach for the identification of faecal pollution sources on an Auckland urban beach. New Zealand Journal of Marine and Freshwater Research, 49, pp 333-345.
Witty M, Nickels J, Lisa J & Tiedemann J (2009): Ecology, DNA, and the Future of Microbial Source Tracking. Water Air and Soil Pollution, 201, pp 219-232.