because (a) a single marker can only be used to identify one particular type of fecal
source and (b) it is unlikely that any one marker will ever be found to be 100 %
sensitive and specific for a particular source in all environments. As a result, a
multivariate “toolbox” approach to MST has been advocated where multiple
markers are measured to obtain a stronger weight of evidence for a particular
source [83]. More recently, an emerging extension of this concept has been to
simply characterize the entire microbial community in a water sample and determine any potential similarities with fecal microbial communities from suspected
sources [84, 85].
Microbial communities are complex assemblages of populations representing
hundreds to thousands of microbial taxa, which is the primary attraction of using
community analysis in MST. These taxa can not only be considered as numerous
potential individual markers, but whole assemblages of taxa may also be identified
that provide higher resolution in discerning sources of fecal pollution than any
single indicator or marker alone. The underlying rationale is based on the evidence
that different host species typically contain measurably and consistently different
microbial communities in the gut and feces, which differ from environmental
communities [86, 87]. Therefore, when fecal contamination occurs, the shift in
the presence and abundance of microbial taxa can be detected and used to help
discern the source. The use of community analysis in MST is a very recent
development and remains an active research area; its use should not be considered
established or routine. This is largely because methods to characterize whole
microbial communities without dependence on laboratory culture (which typically
fails to detect >90 % of microbial taxa) are a new and novel development. As a
result, the methods by which microbial communities can be characterized are
rapidly evolving, and at least three different approaches have been used recently
for MST. Cao et al. [88] employed terminal restriction fragment length polymorphism (TRFLP), a community-scale genetic “fingerprinting” approach, to identify
human sources of contamination. Dubinsky et al. [89] used the PhyloChip microarray, which can simultaneously assay a sample against numerous genetic probes
for known microbial taxa, to discern animal and human sources of fecal contamination in marine waters. Finally, Unno et al. [84] and McLellan et al. [86] both
identified sources of fecal contamination in river and lake systems, respectively,
using next-generation pyrosequencing, which is quickly becoming the gold standard for microbial community analysis [90, 91].
Most recently, Cao et al. [92] simultaneously employed all three of these
methods to identify fecal sources present in the SIPP challenge samples to evaluate
the performance of evolving community analysis methods. All three approaches
correctly identified dominant fecal sources in >90 % of the samples and rarely
detected a source that was not present, illustrating the potential power of community analysis method. This study also identified two other benefits of this approach
to conventional MST methods. Firstly, community analysis has the potential to
discern different sources of the same host-associated marker. For example, all three
methods were able to distinguish septage from human feces and sewage. Secondly,
community analysis can be used to identify sources for which no suitable marker
Microbial Source Tracking: Advances in Research and a Guide to Application
277
source and (b) it is unlikely that any one marker will ever be found to be 100 %
sensitive and specific for a particular source in all environments. As a result, a
multivariate “toolbox” approach to MST has been advocated where multiple
markers are measured to obtain a stronger weight of evidence for a particular
source [83]. More recently, an emerging extension of this concept has been to
simply characterize the entire microbial community in a water sample and determine any potential similarities with fecal microbial communities from suspected
sources [84, 85].
Microbial communities are complex assemblages of populations representing
hundreds to thousands of microbial taxa, which is the primary attraction of using
community analysis in MST. These taxa can not only be considered as numerous
potential individual markers, but whole assemblages of taxa may also be identified
that provide higher resolution in discerning sources of fecal pollution than any
single indicator or marker alone. The underlying rationale is based on the evidence
that different host species typically contain measurably and consistently different
microbial communities in the gut and feces, which differ from environmental
communities [86, 87]. Therefore, when fecal contamination occurs, the shift in
the presence and abundance of microbial taxa can be detected and used to help
discern the source. The use of community analysis in MST is a very recent
development and remains an active research area; its use should not be considered
established or routine. This is largely because methods to characterize whole
microbial communities without dependence on laboratory culture (which typically
fails to detect >90 % of microbial taxa) are a new and novel development. As a
result, the methods by which microbial communities can be characterized are
rapidly evolving, and at least three different approaches have been used recently
for MST. Cao et al. [88] employed terminal restriction fragment length polymorphism (TRFLP), a community-scale genetic “fingerprinting” approach, to identify
human sources of contamination. Dubinsky et al. [89] used the PhyloChip microarray, which can simultaneously assay a sample against numerous genetic probes
for known microbial taxa, to discern animal and human sources of fecal contamination in marine waters. Finally, Unno et al. [84] and McLellan et al. [86] both
identified sources of fecal contamination in river and lake systems, respectively,
using next-generation pyrosequencing, which is quickly becoming the gold standard for microbial community analysis [90, 91].
Most recently, Cao et al. [92] simultaneously employed all three of these
methods to identify fecal sources present in the SIPP challenge samples to evaluate
the performance of evolving community analysis methods. All three approaches
correctly identified dominant fecal sources in >90 % of the samples and rarely
detected a source that was not present, illustrating the potential power of community analysis method. This study also identified two other benefits of this approach
to conventional MST methods. Firstly, community analysis has the potential to
discern different sources of the same host-associated marker. For example, all three
methods were able to distinguish septage from human feces and sewage. Secondly,
community analysis can be used to identify sources for which no suitable marker
Microbial Source Tracking: Advances in Research and a Guide to Application
277