CSIRO PUBLISHING

Marine and Freshwater Research

Review
http://dx.doi.org/10.1071/MF15361

Using environmental (e)DNA sequencing for aquatic

biodiversity surveys: a beginner’s guide

Jennifer L. A. Shaw A,B, Laura Weyrich A and Alan Cooper A

A
Australian Centre for Ancient DNA (ACAD), University of Adelaide, Darling Building,
North Terrace, Adelaide, SA 5000, Australia.
B
Corresponding author. Present address: Commonwealth Scientific and Industrial
Research Organisation, Land and Water Flagship, Prescott Building, Waite Road,
Urrbrae, SA 5064, Australia. Email: jennifer.shaw@csiro.au

Abstract. Biological surveys are needed to monitor and assess the health of ecosystems and the species within them.
However, morphology-based biodiversity surveys can be invasive, time consuming and financially expensive. Recently,
environmental (e)DNA sequencing has been demonstrated as a potential alternative to morphological-based surveys
because it enables the rapid and inexpensive detection of multiple taxa from DNA present in the environment. Numerous
studies have shown that eDNA-based biodiversity surveys can provide considerable information about aquatic ecosystem
function and health. Therefore, this molecular method has the potential to improve how current aquatic biological surveys
are conducted. Currently, most eDNA literature is aimed at an audience with a moderate to advanced knowledge of DNA
sequencing, creating a barrier for many ecologists who lack DNA sequencing expertise but wish to apply such methods to
their research. The aim of this review is to provide guidance to non-geneticists regarding sequencing eDNA for aquatic
biodiversity surveys and to highlight the requirements that need to be considered before the technique can be effectively
incorporated into biomonitoring programs. Specifically, we provide details and recommendations on some of the major
principles, from sample collection to bioinformatic analyses. For those areas where specific recommendations cannot be
given, we have provided references to suitable literature.

Additional keywords: metabarcoding, metagenomics, methods, monitoring, NGS, summary.

Received 21 September 2015, accepted 10 November 2015, published online 1 March 2016

Introduction
Aquatic systems are under mounting pressure from climatic
changes, pollution and overexploitation (Dudgeon et al. 2006;
Crain et al. 2009). As a result, the health of both marine and
freshwater ecosystems has declined globally, in terms of
biodiversity loss (Butchart et al. 2010), species invasions
(Strayer et al. 2006) and water quality (Foley et al. 2005). In
order to prevent further declines, and perhaps improve the health
and quality of these ecosystems in the future, regular monitoring
and management is required. Currently, there are several
commonly used biological monitoring methods for aquatic
systems, each designed for a specific group of organisms. For
example, monitoring small vertebrates or invertebrate species
often involves invasive capture-based surveys, such as netting,
whereas plant surveys are usually observational. Populations of
larger vertebrates may be surveyed using a combination of both
observational and capture methods (Karr and Chu 1999),
whereas surveys of microscopic biota, such as bacteria, algae,
fungi, archea and viruses, may include a variety of laboratory
protocols, such as microscopy, heterotrophic plate counts, flow
cytometry (Hammes et al. 2008), denaturing gradient gel
electrophoresis (Muyzer et al. 1993) or quantitative polymerase
chain reaction (PCR) analysis (Arya et al. 2005). Although each
of these survey methods is informative when applied to the
appropriate taxa of interest, there are associated biases and
limitations. For example, the invasive nature of capture-based
sampling can damage ecosystems and increase predation risk to
organisms (Resources Inventory Committee 1997), and difficul-
ties associated with identifying cryptic species and juvenile life
stages often confound traditional morphological identification-
based methodologies. Further, a continuous decline in taxonomic
expertise and the non-standardised skill levels of different
taxonomists can limit surveys and introduce bias. In addition,
because biological surveys of multiple trophic levels are often
required for effective assessments of aquatic system health,
multiple methods and ecologists specialising in a diverse range
of organisms would be needed, further inflating financial costs.
Because most monitoring programs are financially limited, the
consequence is usually to scale back the extent of the survey,
such as to reduce the number of locations surveyed or samples
collected.
Molecular-based, environmental (e)DNA surveys are a
potential alternative to traditional morphological-based bio-
diversity surveys. ‘eDNA’ simply refers to the DNA that is present

Journal compilation Ó CSIRO 2016 www.publish.csiro.au/journals/mfr

B Marine and Freshwater Research
in an environmental sample, such as water, sediment, soil or
faeces (Taberlet et al. 2012), and sequencing this eDNA enables
identification of taxa from all life stages without morphological
identification. To elaborate, larger organisms, such as
mammals, birds, fish and reptiles, leave DNA traces within
environments in the form of shedding skin, excretion and
releasing gametes during reproduction, and DNA from smaller
organisms can also be detected (e.g. bacteria, plankton etc).
Using eDNA in this way provides a non-invasive and rapid way
to identify many organisms from an environmental sample.
Using eDNA to detect organisms can refer to two different
methods: (1) targeted sequencing, where a specific informative
gene region is targeted and sequenced from within the DNA
mixtures present in a sample; and (2) shotgun sequencing, where
all DNA fragments within a sample are sequenced regardless of
which gene they originate from. Shotgun sequencing may be
useful for studies where there are multiple genes of interest, such
as functional studies (Venter et al. 2004). However, because
shotgun sequencing is non-discriminatory, a large majority of
the sequence data may be uninformative. Targeted eDNA
approaches can be further broken down into two main areas of
research: (1) single species targeted eDNA detection, often used
to detect a particular invasive or threatened species; and (2)
multispecies targeted eDNA sequencing (also known as meta-
barcoding), which is used to obtain information about the
biodiversity of an environment. Single species eDNA detection
describes targeting a DNA sequence (also known as a genetic
marker) that is unique to a particular organism from within an
eDNA mixture. Single species targeted approaches have been
experimentally validated for several important organisms and
can provide semiquantitative abundance estimates (Laramie
et al. 2015; Sigsgaard et al. 2015; Spear et al. 2015). However,
because this approach is limited to detecting a single organism
per reaction, it is not efficient when rapid identification of entire
biological communities is required. In contrast, multispecies
eDNA sequencing (metabarcoding) enables entire groups of
organisms to be detected, such as vertebrates, fungi, plants or
bacteria. It is achieved by targeting a less-specific genetic
marker that comprises of two evolutionarily conserved DNA
sequences (the target regions) interspaced by a highly variable
species-specific sequence. This combination of conserved
regions flanking a variable region makes multispecies identifi-
cation possible in a single sequencing run, and therefore enables
community biodiversity of an ecosystem to be characterised.
Multispecies eDNA sequencing was originally applied to
bacterial communities (Chakravorty et al. 2007; Petrosino et al.
2009) but has since been extended to survey other taxonomic
groupings, such as eukaryotic diversity (Chariton et al. 2010;
Bik et al. 2012), vertebrates (Kelly et al. 2014; Thomsen et al.
2012a), invertebrates (Hajibabaei et al. 2011; Yu et al. 2012),
fungi (Bellemain et al. 2010; Lim et al. 2010), plants (Burgess
et al. 2011; Yoccoz et al. 2012) and viruses (Finkbeiner et al.
2008; Roossinck 2012). At present, the method has been applied
to a range of aquatic sample types, including marine ecosystems
(Fonseca et al. 2010; Kimes et al. 2013), freshwater lakes and
rivers (Sweeney et al. 2011; Drury et al. 2013), soils and
sediments (Fierer et al. 2012) and drinking water distribution
systems (Douterelo et al. 2013; Shaw et al. 2014). These studies
have demonstrated that this method enables the detection of
J. A. Shaw et al.
previously unprecedented biodiversity and can provide useful
information about ecosystem functionality and health. Further,
because a large number of samples can be processed in parallel on
a single sequencing run and multiple species can be detected and
identified from each sample, this approach is also comparably
inexpensive and rapid compared with morphological-based
methods (Baird and Hajibabaei 2012; Taberlet et al. 2012;
Ji et al. 2013).
The benefits of metabarcoding from eDNA are clear, but
many ecologists lack the expertise needed to optimise and
validate the method for their field of research. Further, most
eDNA literature assumes that readers have a firm understanding
of genetic laboratory protocols, and many previous reviews have
only discussed the possibilities and benefits of eDNA sequencing
without explaining the methodological approaches directly. As a
result, there is a lack of easily digestible published information
detailing all the methodological steps for eDNA biodiversity
surveys. This creates a barrier for many scientists who lack a
strong genetic background but want to apply eDNA approaches.
This review serves to provide an overview of eDNA survey
procedures and aims to make this method more accessible to non-
geneticists, allowing this approach to be more widely applied in
aquatic ecosystems research. Specifically, we provide details and
recommendations on some of the major principles, from sample
collection to bioinformatic analyses, which require careful
consideration before eDNA sequencing can be effectively used.
Although this review focuses on aquatic environments, the
underlying concepts also apply to other environments.
Overview of eDNA sequencing for biodiversity surveys
eDNA-based biodiversity surveys are comprised of five key
steps (Fig. 1). Each step of the methodology should be optimised
for the scientific question, the target taxa group and the inherent
limitations associated with the sample type. The initial collec-
tion, processing and DNA extraction from the sample are all
crucial steps for ensuring representative diversity is obtained. In
the first part of the review we provide recommendations on the
most efficient and effective sampling approaches, and then
discuss the advantages and disadvantages of the most commonly
used DNA concentration and extraction techniques.
After extraction, the target DNA in each sample is amplified
by PCR (i.e. many identical DNA copies are made) using
specially designed ‘tagged’ primers. The tagged primers bind
to the target gene region (i.e. genetic marker) that is conserved
across the taxon group of interest and the gene region is
amplified across both the conserved and variable (i.e. species-
specific) sections. During this process, a unique sample-specific
tag is also incorporated into every amplified DNA sequence in
that sample. This step enables multiple samples to be later
combined and sequenced together, greatly reducing the through-
put time and cost of sample processing. A brief description of
PCR-amplification and information regarding the most commonly
used genetic markers is provided later in the review.
Following PCR amplification, the samples are cleaned to
remove all small non-target fragments left over from the PCR.
Failure to remove the small non-target sequences can result in
inefficient sequencing. After PCR, the tagged samples can be
pooled together at equal molar concentrations to make a ‘DNA
A guide to eDNA-based biodiversity surveys Marine and Freshwater Research C

1. Sample collection 2. DNA extraction

Target gene
region within
the template
DNA

1st cycle 2nd cycle 3rd cycle

3. PCR amplification of
5. Bioinformatic analysis specific loci with
tagged primers

NGS
PLATFORM

4. NGS sequencing

Fig. 1. Basic workflow of a targeted environmental (e)DNA sequencing biodiversity survey. (1) Environmental samples are collected in
sterile containers; (2) DNA is concentrated and extracted; (3) the target gene region (genetic marker) is amplified by polymerase chain reaction
(PCR) using primers uniquely tagged for each sample; (4) multiple samples are pooled together for sequencing on a next-generation
sequencing (NGS) platform; and (5) after sequencing, the sequences are split into separate sample files and the raw data is processed and
analysed using various computer packages. Images 1, 2a, 3 and 4 are the authors’ own. Image 5 sourced from www.freeimages.co.uk
(reproduced with permission).

library’. The DNA library is then sequenced using a next-
generation sequencing (NGS) platform. The benefits and limita-
tions of the (currently) most commonly used NGS platforms
(454 Life Sciences, Roche, Branford, CT, USA; Ion Torrent
PGM, Thermo Fisher, Carlsbad, CA, USA; and Illumina MiSeq,
Illumina, San Diego, CA, USA) are reviewed herein, and we
include information on how to design and construct tagged
sequencing primers for generating DNA libraries on NGS
platforms.
The final step of an eDNA survey is data processing and
visualisation of results. Currently, there are numerous software
packages available, including some specifically developed for
users with little programming knowledge and others that enable
more experienced users to customise pipelines. Each step within
the bioinformatic analyses requires user-specified parameters
that can markedly influence results if not chosen appropriately;
therefore, it is imperative to have a general understanding of
bioinformatic processes to enable robust interpretation of the
data. Some of the key bioinformatic steps are detailed in this
review.
Contamination considerations
First and foremost, contamination should always be considered
as a constant risk and minimised where possible when con-
ducting any DNA sequencing study. Contamination can be
introduced into eDNA datasets at multiple stages, the most
critical being during sample collection, DNA extraction and
PCR set-up, as these stages are performed when the DNA is in
low concentration (Cooper and Poinar 2000). To minimise the
amount of contaminant DNA, samples should be collected using
sterile, DNA-free storage containers while wearing clean labo-
ratory gloves and minimising handling of both container and
sample (Daniel and van Oorschot 2011).
Sample preparation and DNA extractions should be performed
in a clean environment using DNA-free equipment. A 70%
ethanol solution is suitable for general cleaning of the laboratory
(e.g. door handles, shelves, sinks); however, laboratory benches
and equipment should be decontaminated more thoroughly using
a 1–3% sodium hypochlorite solution (ideally left in contact with
the surface for 15 min) and subjected to ultraviolet (UV) radiation
D Marine and Freshwater Research
where possible (Gefrides et al. 2010; Fraise et al. 2013). It is
imperative that extraction blank controls (where the sample is
omitted or replaced with nuclease-free water) are extracted at the
same time as the samples (Salter et al. 2014). The extraction blank
controls should be carried through the entire process and
sequenced alongside the samples to monitor any taxa introduced
by laboratory contamination.
Further, specific pre- and post-PCR zones within the laboratory
should be designated (ideally in separate rooms). This is important
to reduce cross-contamination from high-copy, post-PCR products
from previous projects (Kwok and Higuchi 1989). Ideally, move-
ment of all equipment and personnel should be restricted between
the two laboratory areas to a one-way direction, following the
DNA concentration gradient. Full decontamination of all labora-
tory surfaces (and any equipment used) should be performed
before and after each use.
Sample collection
Given the heterogeneous nature of ecosystems, it is difficult to
predict levels of sampling sufficient to capture representative
community diversity. Currently, there are no standards regard-
ing the minimal representative sampling volume needed to
capture the complete community diversity present in an aquatic
ecosystem. Previous eDNA studies have found that sample
volumes ranging from as small as 15 mL (Ficetola et al. 2008) up
to 100 L (Gomez-Alvarez et al. 2012) can both yield diverse
biological communities and a greater number of taxa compared
with conventional methods. For example, a study by Thomsen
et al. (2012a) successfully detected a larger number of marine
fish using only thirty 50-mL samples of seawater pooled into
a 1.5-L composite sample compared with conventional (non-
molecular) survey techniques.
The sample volume required to identify representative bio-
logical communities is likely to be dependent on target taxon
abundance, total biomass, community diversity and the proper-
ties of the sample media. Kelly et al. (2014) found that the
abundance and biomass of the taxon of interest is highly
correlated with the number of DNA sequences detected. For
that reason, samples such as eutrophic waters, sewage effluent
and wastewater, which may contain millions of algal, bacterial
or viral cells per millilitre (Gerardi 2006; Burkholder 2001) may
only require a small sample volume to characterise microbial
community diversity. Further, a smaller volume may also be
sufficient for sample types taken from extreme environments,
such as highly acidic or saline materials, which may have low
diversity due to the dominance of a few highly specialised taxa.
Conversely, the DNA of comparatively less abundant organ-
isms, such as vertebrates, may be more dilute and heterogeneous
in a given environment, and larger sample volumes may be
required as a result. As with traditional morphological-based
surveys, there are ways to determine whether the level of
sampling effort is sufficient. For example, by using permuted
accumulation and rarefaction curves, where the number of
individuals is substituted for the number of sequences associated
with the target taxa (Fig. 2). The optimal sample volume will
ultimately depend on the biological properties of the environ-
ment; it is recommended that pilot studies be performed unless a
priori knowledge of the area is available.
J. A. Shaw et al.
Previous eDNA studies have suggested that the number of
samples taken per site is more important than sample volume.
Andersen et al. (2012) found that soil samples taken from within
large zoo enclosures (40  40 m) accurately represented the
megafaunal diversity present regardless of soil sample volume
(6.5–32.5 g soil), and that large vertebrate taxa were infrequently
detected within spatial replicates throughout a sample site,
highlighting the importance of replication when working with
heterogeneously distributed DNA. Similarly, a study that mea-
sured fish taxon biomass in a controlled marine mesocosm found
that diversity was higher and more representative of the true
diversity when three 1-L samples were assessed compared with
one 15-L sample (Kelly et al. 2014). Therefore, we recommend
that multiple spatial replicates of smaller volumes are collected
in favour of fewer samples of larger volume. Replicates should
ideally include experimental replicates (multiple samples taken
per site) and technical replicates (multiple extractions complet-
ed on a single sample); however, the level of replication is often
limited by time and financial constraints associated with the
survey, and studies have shown that optimal levels of replication
are highly dependent on the detection probability of the taxa
(Ficetola et al. 2015).
A further consideration for eDNA biodiversity surveys is the
persistence of DNA in the environment. A targeted species-
specific eDNA study by Dejean et al. (2011) found that DNA of
vertebrate taxa (frogs and fish) could be detected for 1 month
after the organism was removed from an experimental fresh-
water pond. Similarly, Ficetola et al. (2008) and Thomsen et al.
(2012b) detected the DNA of vertebrate taxa for up to 2 weeks
after their removal from a controlled setting. This has vast
Cut off value
(number of sequences per sample)
Number of
OTUs
Number of sequences or
number of samples
Fig. 2. Rarefaction curves representing taxa (operational taxonomic unit,
OTU) count increasing as number of total sequences or samples increases.
Rarefaction curves represent the mean of repeated resampling of all
sequences. A curve that plateaus (dashed line) indicates that most or all of
the diversity has been characterised. A steep curve that does not plateau (solid
line) indicates that more sampling or deeper sequencing is needed. In order to
assign an even sequencing depth across all samples an arbitrary cut off value
may be assigned, where the sequences from each sample are randomly sub-
sampled up to a specific number of total sequences. However, this is only
appropriate if all samples are reasonably well characterised (dashed lines).
A guide to eDNA-based biodiversity surveys
implications for eDNA surveys of fast-flowing aquatic environ-
ments, because traces of an organism’s DNA could be
sequenced further downstream in uninhabited areas, thus giving
an inaccurate representation of the community. In addition to
this, a recent study by Merkes et al. (2014) found that eDNA has
the potential to be geographically transported by secondary
vectors, such as the faeces of predators and vessels such as
boats. Therefore, these factors should be taken into consideration
when interpreting eDNA survey data. However, the persistence
time of DNA is highly dependent on variables such as tempera-
ture, oxygen levels, pH, UV intensity and bacterial activity
(Burger et al. 1999; Strickler et al. 2015). Therefore, in natural
environments, the persistence time may be much less than in
controlled settings, which may have unnaturally high concentra-
tions of organism DNA.
Sample processing and DNA extraction
Aquatic eDNA surveys usually require sampling volumes of
water that are too large for most extraction protocols. Therefore,
samples may need to be concentrated before DNA extraction.
Filtration and centrifugation are the most commonly used con-
centration methods, but both have associated limitations. For
example, samples containing large amounts of particulate
matter can obstruct filter pores during filtration, whereas
centrifugation may not be able to reach the high speeds required
to pellet small cells or extracellular DNA. Therefore, centrifu-
gation may be more appropriate for samples with large amounts
of particulate matter, and filtration may be more suitable for
concentrating clearer water samples. Alternatively, a combina-
tion of both methods could be used. For example, following
centrifugation, DNA can be extracted directly from the pellet
and the supernatant retained and subjected to filtration to capture
additional DNA. Centrifugation speed can be optimised
according to target taxa, with high g centrifugation (10 000g for
15 min at 48C) effectively concentrating large bacterial cells
(Hoefel et al. 2003), whereas alternative methods are required to
pellet smaller viruses (Lewis and Metcalf 1988). Filtration has
been more commonly used in aquatic system eDNA studies.
When sequencing marine fish DNA from 1.5-L seawater sam-
ples, Thomsen et al. (2012a) filtered the seawater through a
0.47-mm membrane to concentrate the DNA before extracting
the DNA directly from the filter membrane. Douterelo et al.
(2013) and Hwang et al. (2012) both used a 0.22-mm membranes
to filter 1 and 10 L of drinking water respectively for their
studies on bacterial communities. To prevent particulate matter
obstructing filter pores and optimise capture of eDNA, Turner
et al. (2014) used multiple filter sizes to fractionate the sus-
pended particulate matter and the eDNA from water collected
from the Great Lakes. They found eDNA was most abundant
when using 0.2-mm pore sizes, and suspended particulate matter
was most abundant when a 100-mm pore size were used. Fur-
thermore, this sequential ‘stacking’ of a range of filters has the
added benefit of fractionating the sample into different phases,
which may be beneficial for separating intracellular DNA within
intact cells from degraded extracellular DNA.
The often large number of samples in eDNA surveys projects
creates a bottleneck at the extraction stage. Therefore, stream-
lined, standardised, high-throughput extraction kits are useful
Marine and Freshwater Research E
for speeding up processing times. Extraction kits, such as the Mo
Bio (Mo Bio laboratories, Carlsbad, CA, USA) or Qiagen
(Hilden, Germany) range, have been specifically designed to
process difficult environmental samples, including those that
contain high amounts of humic acids and other PCR inhibitors
(i.e. sediments or wastewater samples; see manufacturer web-
sites), in a reasonably rapid manner. However, despite the
obvious benefits of using manufactured extraction kits, studies
have shown that the DNA yield of commercial kits is generally
lower than specially developed ‘in-house’ extraction techniques
(Hurt et al. 2001; Martin-Laurent et al. 2001; Mahmoudi et al.
2011), although DNA quality is arguably more important than
DNA quantity.
Different DNA extraction protocols can produce vastly
different community profiles. For example, Deiner et al.
(2015) found significantly different species detection rates
ranging from 0 to 33% across different extraction protocols.
These biases will likely be more significant in phenotypically
diverse communities. For example, in microbial communities,
fragile cells will lyse more easily than spore or cyst-forming
taxa. If lysis is too gentle, then DNA from robust species will not
be extracted; yet, if lysis is too vigorous, the DNA from more
fragile species could become overly sheared. Both factors may
skew the resulting community data. Therefore, we recommend
that various extraction protocols are tested and optimised for
specific taxa groups and sample types, and that a standardised
extraction method is adopted across all samples.
Many eDNA extraction protocols for biodiversity studies
feature a mechanical lysis step, such as bead beating, in
the presence of a lysis buffer. Lysis buffer is a solution used
for the purpose of lysing open the cells and minimises DNA
damage once the DNA is released. Most lysis buffers contain
salts (e.g. Tris-HCl or EDTA) to regulate the acidity and
osmolality of the extraction solution, whereas detergents (e.g.
Triton-X or sodium dodecyl sulphate (SDS)) are added to break up
membrane structures. The choice of buffer and detergent depends
on sample type and target organisms; for example, animal,
bacterial and yeast cells all have differing requirements for
optimal lysis because of the presence or absence of a cell wall.
Other important considerations for optimal DNA extraction
include temperature, duration of lysis and mechanical lysis choice.
There are many choices of bead for bead-based mechanical
lysis of environmental cells. Garnet beads are non-uniform in
size and shape and have sharp edges; this variation can help
homogenise and break apart sediments while shearing open the
cells within the sample. However, garnet beads are not suitable
for prolonged periods of bead beating because they can easily
break apart and lose the ability to lyse cells effectively. Ceramic,
glass and metal beads are more robust and therefore suitable for
longer bead beating periods, which may be needed to lyse more
resilient cysts and spores. However, it must be noted that
extended bead beating can result in heat damage to extracellular
DNA, leading to lower yields or sheared DNA (von Atzigen and
Kennedy 2011). Therefore, when the target group of interest
contains species of varying fragility, it may be best to use two
separate extraction protocols (more and less intensive) per
sample and combine the extractions at a later stage in the
process. Von Atzigen and Kennedy (2011) recommend a bead
F Marine and Freshwater Research J. A. Shaw et al.

(a) Platform-specific Sample- Gene specific

adapter specific tag primer

Sequence of interest

Gene Sample- Platform-specific

specific specific tag adapter
primer

(b) Linker Gene specific

primer

PCR 1
Sequence of interest (e.g. 30 cycles)

Gene Linker
specific
Platform-specific Sample-
primer
adapter specific tag Linker
Gene specific
primer PCR 2
(e.g. 5 cycles)
Sequence of interest
Gene
specific
primer
Linker Sample- Platform-specific
specific tag adapter

Fig. 3. Polymerase chain reaction (PCR) primer construct for multispecies environmental (e)DNA libraries. (a) Tagged fusion
primer construct requiring a single PCR amplification and (b) a two-step PCR approach where the first PCR amplifies the target
gene region and the second PCR adds an interchangeable sample-specific tag and platform-specific adaptor.

beating speed of 4000 rpm for 45 s as a standard starting point
when extracting DNA from complex environmental communi-
ties. This speed and the length of time should be increased when
targeting taxa with more robust cell walls, such as spores or
cysts, and reduced when dealing with more fragile cells. The
speed and time required will also vary according to the bead type
and size (von Atzigen and Kennedy 2011).
PCR amplification, primer design and genetic marker choice
eDNA-based biological surveys involve amplifying and then
sequencing an appropriate gene region or ‘genetic marker’ to
obtain informative DNA sequences from a target taxon group.
An ideal genetic marker can be defined as a short species- or
genera-specific genomic region flanked on either side by a
genomic region that is conserved across multiple target taxa.
Primer sets are designed to be complementary to the conserved
region, causing the molecule to bind to the DNA fragment and
enabling PCR amplification through the variable (species-
specific) interior section. It is important that primer sets are
equally complementary to all taxa within the target group and
therefore amplify all DNA sequences with similar efficiency to
avoid taxonomic bias (Polz and Cavanaugh 1998). Unique
sample specific ‘tags’ (sometimes referred to as an index or
barcoded tag) are also incorporated into the primer design to
enable discrimination between sequences from different
samples (Fig. 3). Tags are usually ,6–12 bp in length and must
differ from other sample tags by at least two base pairs so that no
single sequencing error can result in misidentification (Binladen
et al. 2007; Shokralla et al. 2012).
There are already a range of commonly cited genetic markers
for eDNA-based biodiversity surveys (Table 1), including the
16S genetic marker for bacteria, the 18S marker for eukaryotes,
and the CO1 and 12S marker for vertebrates. By default the more
commonly used genetic markers tend to have larger sequence
databases, providing broad taxonomic reference collections to
match sequences against. Genes that are highly conserved across
taxonomic space tend to be hampered with a reduction in
phylogenetic resolution for species identification. For example,
18S rDNA, which has been successfully used for eukaryote
communities, and 16S rDNA, which has been similarly used for
bacterial communities, are both conserved across a wide range
of phyla but often lack the resolution to identify sequences
beyond the family or genus level (Creer et al. 2010; Bik et al.
2012). Therefore, careful selection of an appropriate primer set
is crucial for balancing the widest taxonomic coverage with the
best phylogenetic resolution and minimal taxonomic bias.
Despite the existence of these commonly used genetic
markers, it is important that we do not become complacent with
genetic marker choice. Discovery and design of new, more-
effective genetic markers could lead to improved eDNA surveys
in future. To facilitate the discovery of optimal gene regions for
the target taxa, several bioinformatic tools have been developed,
A guide to eDNA-based biodiversity surveys
Table 1. List of commonly used targeted multispecies environmental (e)DNA genetic markers for various taxon groups and references citing their use
CO1, cytochrome c oxidase subunit 1; ITS, internal transcribed spacer; rbcL, ribulose bisphosphate carboxylase large chain gene; trnL, chloroplast tRNA gene
Genetic marker Target taxon group
16S rRNA Bacteria
18S rRNA Eukaryotes
12S rRNA Vertebrates
Mitochondrial CO1 Vertebrates and invertebrates
ITS Fungi and algae
Chloroplast rbcL Plants
Chloroplast trnL Plants
such as PRIMER 3 (Rozen and Skaletsky 1999), ecoPrimer
(Riaz et al. 2011) and GENEIOUS (Kearse et al. 2012). These
tools also enable the in silico simulated amplification of target
taxa against potential primer sets, enabling the amplification
efficiency and genetic resolution of the primers to be studied
before purchase. However, it is important to note that in silico
amplification may not always accurately reflect in vivo applica-
tions (Clarke et al. 2014), and primers should therefore be
experimentally validated before use. Currently, the adoption
of new multispecies genetic markers in eDNA studies is hin-
dered by the limited reference sequences available for some
genomic regions within genetic repositories, and this should be
considered during experimental design.
NGS: primer construct and NGS platform choice
One of the major benefits of eDNA sequencing surveys is the
ability to sequence many samples in parallel, greatly reducing
the throughput time and cost of biomonitoring. The incorpo-
ration of sample-specific tags within the primer set construct
enables many samples to be physically combined and analysed
in parallel, and subsequently separated during data processing.
In addition to the sample-specific tag, a sequencing platform
adaptor sequence must also be incorporated into the design
(Fig. 3). The platform adaptor sequence enables the NGS plat-
form to recognise and sequence the DNA fragments. The
adaptor sequences can be obtained from the platform manu-
facturer websites, along with other useful instructions regarding
DNA library preparation. Once the amplified DNA sequences
within a sample contain the sample-specific tag and a platform
adaptor sequence, multiple samples can be pooled in equimolar
concentrations to form the ‘DNA library’. The DNA library is
then quantified, diluted to a concentration appropriate for the
NGS platform and sequenced.
There are two main methods for constructing eDNA libraries.
The simplest approach involves a single PCR amplification using
so-called fusion primers, which are composed of the platform
adaptor sequence, the unique sample-specific tag and the
sequence complementary to the target gene region (Fig. 3a).
The second approach involves two PCR amplification steps
(Fig. 3b). The first PCR step simply amplifies the target gene
and adds a short artificial linker sequence; the second PCR step
then uses primers that bind to the linker sequence to add the full-
length platform-specific adaptor and sample-specific tag to the
target sequence (de Cárcer et al. 2011). The second PCR step
should be constrained to just a few cycles (approximately five to
Marine and Freshwater Research G
Citations
Claesson et al. (2010), Chakravorty et al. (2007)
Chariton et al. (2010), Medinger et al. (2010), Bik et al. (2012)
Kelly et al. (2014), Hardy et al. (2011), Riaz et al. (2011)
Hebert et al. (2003), Hajibabaei et al. (2011), Yu et al. (2012)
Nilsson et al. (2008), Schoch et al. (2012)
Hollingsworth et al. (2009)
Hollingsworth et al. (2009), Yoccoz et al. (2012)
seven cycles) because unnecessary amplification can skew the
proportion of sequences within the amplification reaction further
(PCR bias; Polz and Cavanaugh 1998). Each approach has
benefits and limitations. The simplicity of the single PCR
approach minimises PCR bias and contamination risk, but there
are considerable financial costs for generating a large number of
fusion primers each with different sample-specific tags. In the
two-step approach, a sample-specific tag could be incorporated
into both the first and second primer sets, creating a larger number
of sample–tag combinations, which would reduce the costs
associated with obtaining many uniquely tagged primers.
Furthermore, the two-step approach allows versatility in the
choice of NGS platform because the platform-specific adaptor
sequence is incorporated during the second PCR step; therefore,
the same primers can be used on multiple platforms.
Several NGS platforms have been marketed since the first
widely available 454 Life Sciences (Roche) system (Mardis
2008; Shokralla et al. 2012; Shendure and Ji 2008). Currently,
the dominant Illumina (MiSeq, NextSeq) and Life Technologies
(Ion Torrent PGM) systems are restricted to sequencing regions
between 100 and 600 bp in length, albeit at very economic
prices and enormous capacity. The estimated data output of each
platform constrains the number of samples that can be run in
parallel, because a greater number of samples reduces the
number of sequences per sample. For example, at the time of
writing, the Miseq platform could generate 25 million sequences
(on average 15 Gb of data) per run; therefore, a sequencing run
containing 100 samples would yield ,250 000 sequences per
sample. However, NGS platforms undergo regular hardware and
software upgrades; therefore, it is recommended that up-to-
date information is obtained directly from the manufacturer’s
website before making a decision regarding platform choice.
Other NGS platforms are associated with more specialist
requirements, such as the long read lengths available using
454 (Life Sciences), or single molecule sequencing using Pacific
Bioscience’s Sequel system (Pacific Biosciences, Menlo Park,
CA, USA).
Bioinformatic analysis
NGS of eDNA produces extremely large datasets that need to be
carefully processed to avoid misleading conclusions. Several
processing steps are required before the data can be interpreted
and analysed (Fig. 4), and each step can affect the results (Kunin
et al. 2010), making it important to fully understand, justify and
report all parameters used. The first step generally involves
H Marine and Freshwater Research
Raw sequences
1. Separate sequences into samples based
on the sample‐specific tag
Discard
unmatched
sequences
2. Trim tags, and primers from the
sequences
Discard
non‐matching
sequences
3. Remove low quality sequences
Discard
low quality
sequences
4. Cluster similar sequences together
(retain information on number of reads
associated with each cluster)
Discard
singular reads
5. Match OTU cluster sequences to
database to identify taxa
OTU table
Fig. 4. Workflow of bioinformatic processing steps used to analyse
metagenomic data. (OTU, operational taxonomic unit.)
searching for the sample-specific tag within the data file to
separate the sequences into multiple sample files. At this stage,
any sequence reads showing mismatches to the specified sample
tag or conserved sequence portion can be discarded, because the
quality of the target sequence may also be unreliable. Sequences
are then trimmed to remove the conserved portion of the gene
region, the sample-specific tag and platform-specific adaptor
sequence, leaving behind the variable genera- or species-
specific portion of the sequence reads. Singleton sequences (i.e.
sequences appearing in the data only once) are usually removed,
because these are likely to be reads containing sequencing
errors. Further, any reads substantially below the expected
genetic marker length (e.g. 80 bp) are usually discarded.
Following these initial processing steps, the remaining
sequences are filtered to remove low-quality sequences (i.e.
platform software struggled to determine which nucleobase,
A, G, C or T, was present), which are likely to produce spurious
results. Quality is assessed using Phred scores. A Phred score of
between Q20 and Q30, which indicates a probability that a base
has a respective ,1 : 100 or ,1 : 1000 chance of being incorrect,
would be appropriate because the sheer volume of NGS data
means that very conservative quality values can be applied to
limit analyses to only the highest-quality data (Mende et al.
2012; Patel and Jain 2012).
To reduce computational demand and increase processing
speed, many researchers opt to cluster all highly similar
sequences together as a single taxonomic cluster before
J. A. Shaw et al.
taxonomically identifying the sequences. These clusters of
highly similar sequences are known as operational taxonomic
units (OTUs). Low similarity thresholds (i.e. sequences ,90%
similar) will reduce the number of OTUs in need of identifica-
tion at the expense of potentially grouping multiple species
together within a single OTU. High similarity thresholds (i.e.
100% similar) can still reduce computational demand by sub-
suming identical sequences into one OTU, retaining species and
genus information, but single sequence errors will cause the
same species to be clustered into separate OTUs. A clustering
threshold of ,97% is often cited in eDNA literature and was
initially shown to be optimal for distinguishing bacterial species
using a hypervariable portion of the 16S gene (Stackebrandt and
Goebel 1994). However, it must be noted that this use of an
arbitrary fixed clustering threshold is inherently flawed, because
lineages evolve at variable rates and no single cut-off value can
realistically accommodate an entire class or phyla of organisms.
A single clustering threshold will inevitably always be too
relaxed for slow-evolving lineages and too stringent for rapidly
evolving ones (Stackebrandt and Goebel 1994; Sogin et al.
2006; Koeppel and Wu 2013).
In addition to choosing optimal Phred score and clustering
thresholds, the method of clustering is an important parameter
(Bik et al. 2012). Sequences can be clustered several ways,
namely de novo, closed reference and open reference, by many
different algorithms. Because of the ample size of multispecies
eDNA datasets, de novo clustering methods offer a rapid and
practical algorithm choice (Edgar 2010; Ghodsi et al. 2011; Fu
et al. 2012). During de novo clustering, a sequence is drawn out
of the dataset and becomes the centre of a new OTU centroid;
this centroid is then compared with all other sequences remaining
in the dataset pool. Sequences that are within the specified
clustering threshold to the centroid are moved from the dataset
pool to the OTU. The OTU is then closed. These steps
are repeated until no sequences remain in the pool. Therefore,
de novo clustering methods are strongly influenced by the
sequence input order because most algorithms do not re-evaluate
previous OTU centroids as clustering progresses. This can result
in inaccurately formed OTUs, where closely related species are
separated and unrelated species are grouped (Koeppel and Wu
2013). Some de novo algorithms pre-sort sequences by their
abundance and then sort through the sequences in order of the
most abundant sequences first (e.g. Uclust; http://www.drive5.
com/uclust/, accessed January 2015), which prevents erroneous
sequences from forming centroids. Other de novo algorithms
have adopted useful strategies to avoid both fixed clustering
thresholds, and sequence input-order dependency (e.g. Swarm
clustering method; Mahé et al. 2014).
Alternative, clustering methods include closed and open
reference picking. Closed reference clustering algorithms use
taxonomically verified sequences from genetic databases as the
OTU centroids. Sequences within the data pool that are within
the similarity threshold to species within the genetic database
are clustered together. Although this method offers a more
stringent approach to OTU picking, it is limited to the sequences
available within the genetic database, and biases could be
introduced by database incompleteness (Bik et al. 2012). To
overcome this, open reference OTU picking operates by carrying
out closed reference OTU clustering, followed by de novo
A guide to eDNA-based biodiversity surveys
clustering on any remaining sequences outside the specified
threshold.
Finally, the filtered and clustered OTU sequences are com-
pared with a genetic reference database to identify the taxa
present. Either the most representative (abundant) sequence in
an OTU or a consensus of all the sequences within an OTU is
selected to search and match to taxonomically verified
sequences within a genetic database. GenBank is a comprehen-
sive online genetic database (http://www.ncbi.nlm.nih.gov/
genbank, accessed January 2015); however, it is not as exten-
sively curated as other smaller databases and, as such, sequence
matches may be inaccurate or uninformative. More heavily
curated databases, such as Greengenes (bacterial 16S rRNA;
http://greengenes.secondgenome.com, accessed January 2015),
UNITE (fungal internal transcribed spacer (ITS); http://www2.
dpes.gu.se/project/unite/UNITE_intro.htm, accessed January
2015), SILVA (eukaryotic 16S/18S rRNA; http://www.arb-
silva.de/, accessed January 2015) and BOLD (eukaryotic
CO1; http://www.barcodinglife.com/, accessed January 2015)
enable accurate and reliable taxonomic identification for eDNA
data. However, for studies involving a specific group of taxa
(e.g. fish from a specific region), it may be appropriate to select a
subset of GenBank sequences and create a personally curated
local database to increase computational efficiency at this step.
Once taxonomic identities have been assigned to each OTU
cluster, a species abundance table (or OTU table) can be
generated that describes the number of sequences attributed to
each OTU in each sample. An OTU table is the basic starting
point with which data interpretation and visualisation can begin.
Bioinformatic tools
Custom building a set of command line scripts to process eDNA
data is a flexible option that encourages an advanced level of
analytical comprehension. Custom bioinformatics pipelines can
be built using tools such as GALAXY (Blankenberg et al. 2010),
PERL (http://perldoc.perl.org/, accessed January 2015), Bioperl
(http://doc.bioperl.org/, accessed January 2015) and R
(R Foundation for Statistical Computing, Vienna, Austria; see
http://www.R-project.org, accessed January 2015), followed by
programs such as MEGAN (Meta Genome Analyzer; Huson
et al. 2007), EXPLICET (Robertson et al. 2013) and R (see
http://www.R-project.org) to visualise the data. However,
development of customised pipelines requires a moderate to
advanced knowledge of computer scripting and, as a result,
many researchers may opt for online data processing programs,
such as the Ribosome Database Project (RDP; Cole et al. 2005)
and MG-RAST (Meyer et al. 2008). These programs are spe-
cifically developed to allow standardised eDNA data processing
and analysis for scientists with few bioinformatics skills.
However, although these programs are useful, a lack of under-
standing of the processes and key assumptions made during the
processing steps can result in incorrect data interpretation. The
publicly available QIIME software (Caporaso et al. 2010) is a
popular choice for scientists who may not be confident enough
to create their own customised pipelines but who want to
maintain control and flexibility when analysing eDNA data.
QIIME provides a user-friendly, but customisable, workflow
that involves selecting from a set of ready-made scripts and
Marine and Freshwater Research I
parameters in the pipeline. It encourages researchers to process
data themselves by providing guidance for each step of the
process. Within QIIME, the UniFrac software provides useful
diversity estimates based on phylogenetic divergence between
taxa (Lozupone and Knight 2005; Lozupone et al. 2006) and can
be used to compare samples and reveal patterns (Lozupone and
Knight 2008). Regardless of the data processing approaches
used, it is crucial that each of the bioinformatics steps is justified
and selected according to the research question applied. It is
recommended that where possible scientists analyse their own
data to minimise misinterpretation of data, using bioinformatic
tools that suit their individual skill level.
Visualisation and interpretation of eDNA data
An OTU abundance table is the common starting point for data
visualisation and interpretation. From this matrix, the species
detected, community diversity and patterns in community
structure can be examined. However, there are a few important
steps to consider. First, it is important to normalise the number of
sequences across samples, because the number of sequences in
each sample can vary considerably because of pipetting error or
unavoidable sequencing biases. Normalisation can be achieved
by converting the total sequence counts per sample into pro-
portions or percentages. Alternatively, the number of sequences
per sample could be rarefied to a specified value. Rarefaction
randomly selects a specified number of sequences from the
sequence pool in a permutational manner so that each sample
contains an even number of sequences. Rarefaction curves,
plotting the number of OTUs as a function of the number of
sequences, can be viewed to determine whether a cut-off value
(limiting the number of sequences used per sample) is appro-
priate for the dataset (Fig 2); a steep slope indicates that a large
fraction of the species diversity remains to be discovered, a flatter
slope indicates that a reasonable number of individual sequences
have been considered in the rarefaction. Therefore, specifying
a rarefaction cut-off value should be avoided if the curve slope
for any one of the samples is disproportionately steep.
In addition to normalising the data, it may be necessary to
apply transformations to the data before carrying out statistical
analyses to down weight highly abundant taxa or vice versa, as
multispecies eDNA data rarely follow a normal distribution.
There are several software programs that enable informative
visualisation of multispecies eDNA data (e.g. QIIME,
EXPLICET, MEGAN, R). Biodiversity measures, such as the
Shannon–Wiener and Chao diversity indices, may be applied to
each sample to compare species richness and evenness across
samples. Box plots, phylogenetic trees, heat maps, principal
coordinate analysis (PCoA) plots and non-metric multidimen-
sional scaling (nMDS) plots can be used to visualise patterns in
the community structure. These visualisation tools enable data
to be easily interpreted by ecologists that may not have a strong
genetic background. OTU tables can often be exported into
statistical software programs that were not initially designed for
interpretation of eDNA data (e.g. Primer-E; Clarke 1993).
Because most biomonitoring organisations already use multi-
variate analysis software tools such as Primer-E and R, at this
stage eDNA data should be easily transferrable to researchers
from non-genetic backgrounds.
J Marine and Freshwater Research
Sequence read abundance and species richness
Strictly quantitative organism abundance measures based on the
sequence read abundance is not recommended for multispecies
eDNA data. This is due to several reasons. For example, most
organisms contain multiple copies of the target gene in their cells,
and these copy numbers can vary by an order of magnitude across
species (Rooney and Ward 2005; Lee et al. 2009). Further, sev-
eral steps in the eDNA protocol can skew the relative abundance
of a sequence read even further, such as DNA extraction
(DeSantis et al. 2005; Feinstein et al. 2009), primer choice
(Jumpponen 2007; Engelbrektson et al. 2010), PCR bias (Polz
and Cavanaugh 1998) and sequencing biases and errors (Kunin
et al. 2010). However, several studies have found that read counts
are semiquantitative, where differences in the proportional
abundance of sequences for a given species across samples
reflects the actual proportional abundance of that species in the
environment (Porazinska et al. 2009; Kelly et al. 2014; Elbrecht
and Leese 2015). Despite this, the ability to infer species abun-
dance from multispecies eDNA datasets is still regarded as highly
contentious (Deagle et al. 2013), especially for larger eukaryotic
organisms, which may have considerably variable biomasses and
gene copy numbers. In particular, a study by Amend et al. (2010)
examined the relationship between read abundance and biologi-
cal abundance for known quantities and identities of fungi along a
dilution gradient and found that read abundance is approximately
quantitative within species and across samples, but between-
species comparisons are more tenuous. Further, if validation of
proportional species abundances is required, multispecies data
could be verified using single species genetic markers, because
several studies have identified strong positive relationships
between read abundance and population densities using species-
specific targeted eDNA markers (Takahara et al. 2012; Thomsen
et al. 2012b; Pilliod et al. 2013).
Concluding remarks
NGS technologies and bioinformatics programs have provided
revolutionary new ways to assess biodiversity in complex eco-
systems (Binladen et al. 2007). To date, eDNA has been used to
successfully assess community structure and biodiversity in
several aquatic habitats, including marine, lakes, river systems
and drinking water distribution systems. However, in order for
this technique to become a regular biomonitoring tool, impor-
tant concepts and considerations need to be better understood by
ecologists by making DNA-based protocols more transparent
and understandable to non-geneticists. Recent improvements in
sequencing platforms and bioinformatic tools, as well a reduction
in the sequencing time and financial costs of high-throughput
genomics (Logue et al. 2008; Pennisi 2011), offer strong potential
for eDNA approaches to be used as a biodiversity monitoring tool.
This review provides a beginners guide to the practical applica-
tion of multispecies eDNA sequencing to aquatic systems and
highlights several important considerations for integrating this
approach into routine biomonitoring programs.
References
Amend, A. S., Seifert, K. A., and Bruns, T. D. (2010). Quantifying microbial
communities with 454 pyrosequencing: does read abundance count?
J. A. Shaw et al.
Molecular Ecology 19, 5555–5565. doi:10.1111/J.1365-294X.2010.
04898.X
Andersen, K., Bird, K. L., Rasmussen, M., Haile, J., Breuning-madsen, H.,
Kjaer, K. H., Orlando, L., Gilbert, M. T., and Willerslev, E. (2012). Meta-
barcoding of ‘dirt’ DNA from soil reflects vertebrate biodiversity. Molec-
ular Ecology 21, 1966–1979. doi:10.1111/J.1365-294X.2011.05261.X
Arya, M., Shergill, I. S., Williamson, M., Gommersall, L., Arya, N., and
Patel, H. R. (2005). Basic principles of real-time quantitative PCR.
Expert Review of Molecular Diagnostics 5, 209–219. doi:10.1586/
14737159.5.2.209
Baird, D. J., and Hajibabaei, M. (2012). Biomonitoring 2.0: a new paradigm
in ecosystem assessment made possible by next-generation DNA
sequencing. Molecular Ecology 21, 2039–2044. doi:10.1111/J.1365-
294X.2012.05519.X
Bellemain, E., Carlsen, T., and Brochmann, C. (2010). ITS as DNA bar-code
for fungi: an in silico approach reveals potential PCR biases. BMC
Microbiology 10, 189. doi:10.1186/1471-2180-10-189
Bik, H. M., Porazinska, D. L., Creer, S., Caporaso, J. G., Knight, R., and
Thomas, W. K. (2012). Sequencing our way towards understanding
global eukaryotic biodiversity. Trends in Ecology & Evolution 27,
233–243. doi:10.1016/J.TREE.2011.11.010
Binladen, J., Gilbert, M. T. P., Bollback, J. P., Panitz, F., and Bendixen, C.
(2007). The use of coded PCR primers enables high-throughput sequencing
of multiple homolog amplification products by 454 parallel sequencing.
PLoS One 2, e197. doi:10.1371/JOURNAL.PONE.0000197
Blankenberg, D., Von Kuster, G., Coraor, N., Ananda, G., Lazarus, R.,
Mangan, M., Nekrutenko, A., and Taylor, J. (2010). Galaxy: a web-based
genome analysis tool for experimentalists. Current Protocols in Mole-
cular Biology 89(19.10), 19.10.1–19.10.21. doi:10.1002/0471142727.
MB1910S89
Burger, J., Hummel, S., Herrmann, B., and Henke, W. (1999). DNA
preservation: a microsatellite-DNA study on ancient skeletal remains.
Electrophoresis 20, 1722–1728. doi:10.1002/(SICI)1522-2683
(19990101)20:8,1722::AID-ELPS1722.3.0.CO;2-4
Burgess, K. S., Fazekas, A. J., and Kesanakurti, P. R. (2011). Discriminating
plant species in a local temperate flora using the rbcL and matK DNA
barcode. Methods in Ecology and Evolution 2, 333–340. doi:10.1111/
J.2041-210X.2011.00092.X
Burkholder, J. M. (2001). Eutrophication and oligotrophication. In ‘Ency-
clopedia of Biodiversity’. (Ed. S. Levin.) Vol. 2, pp. 649–670.
(Academic Press: New York.)
Butchart, S. H. M., Walpole, M., Collen, B., van Strien, A., Scharlemann,
J. P. W., Almond, R. E. A., Baillie, J. E. M., Bomhard, B., Brown, C.,
Bruno, J., Carpenter, K. E., Carr, G. M., Chanson, J., Chenery, A. M.,
Csirke, J., Davidson, N. C., Dentener, F., Foster, M., Galli, A., Galloway,
J. N., Genovesi, P., Gregory, R. D., Hockings, M., Kapos, V., Lamarque,
J., Leverington, F., Loh, J., McGeoch, M. A., McRae, L., Minasyan, A.,
Hernandez Morcillo, M., Oldfield, T. E. E., Pauly, D., Quader, S.,
Revenga, C., Sauer, J. R., Skolnik, B., Spear, D., Stanwell-smith, D.,
Stuart, S. N., Symes, A., Tierney, M., Tyrell, T. D., and Vie Reg Watson, J.
(2010). Global biodiversity: indicators of recent declines. Science 328,
1164–1168. doi:10.1126/SCIENCE.1187512
Caporaso, J. G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F. D.,
Costello, E. K., Fierer, N., Gonzalez Peña, A., Goodrich, J. K., Gordon,
J. I., Huttley, G. A., Kelley, S. T., Knights, D., Koenig, J. E., Ley, R. E.,
Lozupone, C. A., McDonald, D., Muegge, B. D., Pirrung, M., Reeder, J.,
Sevinsky, J. R., Turnbaugh, P. J., Walters, W. A., Widmann, J.,
Yatsunenko, T., Zaneveld, J., and Knight, R. (2010). QIIME allows
analysis of high-throughput community sequencing data. Nature Methods
7, 335–336. doi:10.1038/NMETH.F.303
Chakravorty, S., Helb, D., Burday, M., Connell, N., and Alland, D. (2007).
A detailed analysis of 16S ribosomal RNA gene segments for the
diagnosis of pathogenic bacteria. Journal of Microbiological Methods
69, 330–339. doi:10.1016/J.MIMET.2007.02.005
A guide to eDNA-based biodiversity surveys
Chariton, A. A., Court, L. N., Hartley, D. M., Colloff, M. J., and Hardy, C. M.
(2010). Ecological assessment of estuarine sediments by pyrosequencing
eukaryotic ribosomal DNA. Frontiers in Ecology and the Environment 8,
233–238. doi:10.1890/090115
Claesson, M., Wang, Q., and O’Sullivan, O. (2010). Comparison of two
next-generation sequencing technologies for resolving highly complex
microbiota composition using tandem variable 16S rRNA gene regions.
Nucleic Acids Research 38, e200. doi:10.1093/NAR/GKQ873
Clarke, K. R. (1993). Non-parametric multivariate analyses of changes in
community structure. Australian Journal of Ecology 18, 117–143.
doi:10.1111/J.1442-9993.1993.TB00438.X
Clarke, L. J., Soubrier, J., Weyrich, L., and Cooper, A. (2014). Environmen-
tal metabarcodes for insects: in silico PCR reveals potential for
taxonomic bias. Molecular Ecology 14, 1160–1170. doi:10.1111/
1755-0998.12265
Cole, J. R., Chai, B., Farris, R. J., Wang, Q., Kulam, S. A., McGarrell, D. M.,
Garrity, G. M., and Tiedje, J. M. (2005). The Ribosomal Database
Project (RDP-II): sequences and tools for high-throughput rRNA analysis.
Nucleic Acids Research 33, D294–D296. doi:10.1093/NAR/GKI038
Cooper, A., and Poinar, H. (2000). Ancient DNA: do it right or not at all.
Science 289, 1139b. doi:10.1126/SCIENCE.289.5482.1139B
Crain, C. C., Halpern, B. S., Beck, M. W., and Kappel, C. V. (2009).
Understanding and managing human threats to the coastal marine
environment. Annals of the New York Academy of Sciences 1162, 39–62.
doi:10.1111/J.1749-6632.2009.04496.X
Creer, S., Fonseca, V. G., Porazinska, D. L., Giblin-davis, R. M., Sung, W.,
Power, D. M., Packer, M., Carvalho, G. R., Blaxter, M. L., Lambshead,
P. J., and Thomas, W. K. (2010). Ultrasequencing of the meiofaunal
biosphere: practice, pitfalls and promises. Molecular Ecology 19, 4–20.
doi:10.1111/J.1365-294X.2009.04473.X
Daniel, R., and van Oorschot, R. A. H. (2011). An investigation of the
presence of DNA on unused laboratory gloves. Forensic Science
International 45, e46.
de Cárcer, D. A., Denman, S. E., McSweeney, C., and Morrison, M. (2011).
Strategy for modular tagged high-throughput amplicon sequencing.
Applied and Environmental Microbiology 77, 6310–6312. doi:10.1128/
AEM.05146-11
Deagle, B. E., Thomas, A. C., Shaffer, A. K., Trites, A. W., and Jarman, S. N.
(2013). Quantifying sequence proportions in a DNA-based diet study
using Ion Torrent amplicon sequencing: which counts count? Molecular
Ecology Resources 13, 620–633. doi:10.1111/1755-0998.12103
Deiner, K., Walser, J., Machler, E., and Altermatt, F. (2015). Choice of
capture and extraction methods affect detection of freshwater biodiver-
sity from environmental DNA. Biological Conservation 183, 53–63.
doi:10.1016/J.BIOCON.2014.11.018
Dejean, T., Valentini, A., Duparc, A., Pellier-Cuit, S., and Pompanon, F.
(2011). Persistence of environmental DNA in freshwater ecosystems.
PLoS One 6, e23398. doi:10.1371/JOURNAL.PONE.0023398
DeSantis, T., Stone, C., Murray, S., Moberg, J., and Andersen, G. (2005).
Rapid quantification and taxonomic classification of environmental
DNA from both prokaryotic and eukaryotic origins using a microarray.
FEMS Microbiology Letters 245, 271–278. doi:10.1016/J.FEMSLE.
2005.03.016
Douterelo, I., Sharpe, R. L., and Boxall, J. B. (2013). Influence of hydraulic
regimes on bacterial community structure and composition in an
experimental drinking water distribution system. Water Research 47,
503–516. doi:10.1016/J.WATRES.2012.09.053
Drury, B., Rosi-Marshall, E., and Kelly, J. J. (2013). Wastewater treatment
effluent reduces abundance and diversity of benthic bacterial communities
in urban and suburban rivers. Applied and Environmental Microbiology
79, 1897. doi:10.1128/AEM.03527-12
Dudgeon, D., Arthington, A. H., Gessner, M. O., Kawabata, Z., Knowler, D. J.,
Leveque, C., Naiman, R. J., Prieur-Richard, A. H., Soto, D., Stiassny, M. L.,
and Sullivan, C. A. (2006). Freshwater biodiversity: importance, status and
Marine and Freshwater Research K
conservation challenges. Biological Reviews of the Cambridge Philo-
sophical Society 81, 163–182. doi:10.1017/S1464793105006950
Edgar, R. C. (2010). Search and clustering orders of magnitude faster than
BLAST. Bioinformatics 26, 2460–2461. doi:10.1093/BIOINFORMAT
ICS/BTQ461
Elbrecht, V., and Leese, F. (2015). Can DNA-based ecosystem assessments
quantify species abundance? Testing primer bias and biomass–sequence
relationships with an innovative metabarcoding protocol. PLoS ONE 10,
e0130324. doi:10.1371/journal.pone.0130324
Engelbrektson, A., Kunin, V., Wrighton, K., Zvenigorodsky, N., Chen, F.,
Ochman, H., and Hugenholtz, P. (2010). Experimental factors affecting
PCR-based estimates of microbial species richness and evenness. The
ISME Journal 4, 642–647. doi:10.1038/ISMEJ.2009.153
Feinstein, L. M., Sul, W. J., and Blackwood, C. B. (2009). Assessment of
bias associated with incomplete extraction of microbial DNA from soil.
Applied and Environmental Microbiology 75, 5428–5433. doi:10.1128/
AEM.00120-09
Ficetola, G. F., Miaud, C., Pompanon, F., and Taberlet, P. (2008). Species
detection using environmental DNA from water samples. Biology
Letters 4, 423–425. doi:10.1098/RSBL.2008.0118
Ficetola, G. F., Pansu, J., Bonin, A., Coissac, E., Giguet-Covex, C.,
De Barba, M., Gielly, L., Lopes, C. M., Boyer, F., Pompanon, F.,
Raye, G., and Taberlet, P. (2015). Replication levels, false presences
and the estimation of the presence/absence from eDNA metabarcoding
data. Molecular Ecology Resources 15, 543–556. doi:10.1111/1755-
0998.12338
Fierer, N., Leff, J. W., Adams, B. J., Nielsen, U. N., Bates, S. T., Lauber, C. L.,
Owens, S., Gilbert, J. A., Wall, D. H., and Caporaso, G. (2012). Cross-
biome metagenomic analyses of soil microbial communities and their
functional attributes. Proceedings of the National Academy of Sciences of
the United States of America 109, 21390–21395. doi:10.1073/PNAS.
1215210110
Finkbeiner, S. R., Allred, A. F., Tarr, P. I., Klein, E. J., Kirkwood, C. D., and
Wang, D. (2008). Metagenomic analysis of human diarrhea: viral
detection and discovery. PLoS Pathogens 4, e1000011. doi:10.1371/
JOURNAL.PPAT.1000011
Foley, J. A., DeFries, R., Asner, G. P., Barford, C., Bonan, G., Carpenter, S.
R., Chapin, F. S., Coe, M. T., Daily, G., Gibbs, H. K., Helkowski, J. H.,
Holloway, T., Howard, E. A., Kucharik, C. J., Monfreda, C., Patz, J. A.,
Prentice, C., Ramankutty, N., and Snyder, P. K. (2005). Global con-
sequences of land use. Science 309, 570–574. doi:10.1126/SCIENCE.
1111772
Fonseca, V. G., Carvalho, G. R., Sung, W., Johnson, H. F., Power, D. M.,
Neill, S. P., Packer, M., Blaxter, M. L., Lambshead, P. J. D., Thomas, W. K.,
and Creer, S. (2010). Second-generation environmental sequencing
unmasks marine metazoan biodiversity. Nature Communications 1, 98.
doi:10.1038/NCOMMS1095
Fraise, A. P., Maillard, J. Y., and Sattar, S. A. (2013). ‘Russell, Hugo &
Ayliffe’s Principles and Practice of Disinfection, Preservation and
Sterilization’, 5th edn. (Wiley-Blackwell: Oxford, UK.) doi:10.1002/
9781118425831.index
Fu, L., Niu, B., Zhu, Z., Wu, S., and Li, W. (2012). CD-HIT: accelerated for
clustering the next-generation sequencing data. Bioinformatics 28,
3150–3152. doi:10.1093/BIOINFORMATICS/BTS565
Gefrides, L. A., Powell, M. C., Donley, M. A., and Kahn, R. (2010). UV
irradiation and autoclave treatment for elimination of contaminating
DNA from laboratory consumables. Forensic Science International:
Genetics 4, 89–94. doi:10.1016/J.FSIGEN.2009.06.008
Gerardi, M. H. (2006). ‘Wastewater Bacteria.’ (Wiley: Hoboken, NJ.)
Ghodsi, M., Liu, B., and Pop, M. (2011). DNACLUST: accurate and
efficient clustering of phylogenetic marker genes. BMC Bioinformatics
12, 271. doi:10.1186/1471-2105-12-271
Gomez-Alvarez, V., Revetta, R. P., and Santo Domingo, J. W. (2012).
Metagenomic analyses of drinking water receiving different disinfection
L Marine and Freshwater Research
treatments. Applied and Environmental Microbiology 78, 6095–6102.
doi:10.1128/AEM.01018-12
Gotelli, N. J., and Colwell, R. K. (2001). Quantifying biodiversity: proce-
dures and pitfalls in the measurement and comparison of species
richness. Ecology Letters 4, 379–391. doi:10.1046/J.1461-0248.2001.
00230.X
Hajibabaei, M., Shokralla, S., Zhou, X., Singer, G. A. C., and Baird, D. J.
(2011). Environmental barcoding: a next-generation sequencing
approach for biomonitoring applications using river benthos. PLoS
One 6, e17497. doi:10.1371/JOURNAL.PONE.0017497
Hammes, F., Berney, M., Wang, Y., Vital, M., Koster, O., and Egli, T.
(2008). Flow-cytometric total bacterial cell counts as a descriptive
microbiological parameter for drinking water treatment processes.
Water Research 42, 269–277. doi:10.1016/J.WATRES.2007.07.009
Hardy, C. M., Adams, M., Jerry, D. R., Court, L. N., Morgan, M. J., and
Hartley, D. M. (2011). DNA barcoding to support conservation: species
identification, genetic structure, and biogeography of fishes in the
Murray–Darling River Basin, Australia. Marine and Freshwater
Research 62, 887–901. doi:10.1071/MF11027
Hebert, P. D. N., Ratnasingham, S., and deWaard, J. R. (2003). Barcoding
animal life: cytochrome c oxidase subunit 1 divergences among closely
related species. Proceedings. Biological Sciences 270, S96–S99.
doi:10.1098/RSBL.2003.0025
Hoefel, D., Grooby, W. L., Monis, P., Andrews, S., and Saint, C. (2003). A
comparative study of carboxyflourescin diacetate and carboxyflourescin
diacetate succinimidyl ester as indicators of bacterial activity. Journal of
Microbiological Methods 52, 379–388. doi:10.1016/S0167-7012(02)
00207-5
Hollingsworth, P. M., Forrest, L. L., Spouge, J. L., Hajibabaei, M., and
Ratnasingham, S. M. van der Bank, M. W. Chase, R. S. Cowan, D. L.
Erickson, A. J. Fazekas, S. W. Graham, K. E. Jamesi, K.-J. Kim, W. J.
Kress, H. Schneider, J. van AlphenStahl, S. C.H. Barrett, C. van den
Berg, D. Bogarin, K. S. Burgess, K. M. Camerono, M. Carine, J.
Chacón, A. Clark, J. J. Clarkson, F. Conrad, D. S. Devey, C. S. Ford,
T. A. J. Hedderson, M. L. Hollingsworth, B. C. Husband, L. J. Kelly,
P. R. Kesanakurti, J. Sung Kim, Y.-D. Kim, R. Lahaye, H.-L. Lee, D. G.
Long, S. Madriñán, O. Maurin, I. Meusnier, S. G. Newmaster, C.-W.
Park, D. M. Percy, G. Petersen, J. E. Richardson, G. A. Salazar, V.
Savolainen, O. Seberg, M. J. Wilkinson, D.-K. Yi, and D. P. Little
(2009). A DNA barcode for land plants. Proceedings of the National
Academy of Sciences of the United States of America 106, 12794–12797.
doi:10.1073/PNAS.0905845106
Hurt, R. A., Qiu, X., Wu, L., Roh, Y., Palumbo, A. V., Tiedje, J. M., and
Zhou, J. (2001). Simultaneous recovery of RNA and DNA from soils and
sediments. Applied and Environmental Microbiology 67, 4495–4503.
doi:10.1128/AEM.67.10.4495-4503.2001
Huson, D. H., Auch, A. F., Qi, J., and Schuster, S. C. (2007). MEGAN
analysis of metagenomic data. Genome Research 17, 377–386.
doi:10.1101/GR.5969107
Hwang, C., Ling, F., Andersen, G. L., LeChevallier, M. W., and Liu, W.
(2012). Microbial community dynamics of an urban drinking water
distribution system subjected to phases of chloramination and chlorina-
tion treatments. Applied and Environmental Microbiology 78, 7856.
doi:10.1128/AEM.01892-12
Ji, Y., Ashton, L., Pedley, S. M., Edwards, D. P., Tang, Y., Nakamura, A.,
Kitching, R., Dolman, P. M., Woodcock, P., Edwards, F. A., Larsen,
T. H., Hsu, W. W., Benedick, S., Hamer, K. C., Wilcove, D. S., Bruce, C.,
Wang, X., Levi, T., Lott, M., Emerson, B. C., and Yu, D. W. (2013).
Reliable, verifiable and efficient monitoring of biodiversity via meta-
barcoding. Ecology Letters 16, 1245–1257. doi:10.1111/ELE.12162
Jumpponen, A. (2007). Soil fungal communities underneath willow cano-
pies on a primary successional glacier forefront: rDNA sequence results
can be affected by primer selection and chimeric data. Microbial
Ecology 53, 233–246. doi:10.1007/S00248-004-0006-X
J. A. Shaw et al.
Karr, J. R., and Chu, E. W. (1999). ‘Restoring Life in Running Waters: Better
Biological Monitoring.’ (Island Press: Washington, DC.)
Kearse, M., Moir, R., Wilson, A., Stones-Havas, S., Cheung, M., Sturrock, S.,
Buxton, S., Cooper, A., Markowitz, S., Duran, C., Thierer, T., Ashton, B.,
Meintjes, P., and Drummond, A. (2012). Geneious Basic: an integrated
and extendable desktop software platform for the organisation and
analysis of sequence data. Bioinformatics 28, 1647–1649. doi:10.1093/
BIOINFORMATICS/BTS199
Kelly, R. P., Port, J. A., Yamahara, K. M., and Crowder, L. B. (2014). Using
environmental DNA to census marine fishes in a large mesocosm. PLoS
One 9, e86175. doi:10.1371/JOURNAL.PONE.0086175
Kimes, N. E., Callaghan, A. V., Aktas, D. F., Smith, W. L., Sunner, J.,
Golding, B., Drozdowska, M., Hazen, T. C., Suflita, J. M., and Morris, P. J.
(2013). Metagenomic analysis and metabolite profiling of deep-sea
sediments from the Gulf of Mexico following the Deepwater Horizon oil
spill. Frontiers in Microbiology 4, 50. doi:10.3389/FMICB.2013.00050
Koeppel, A. F., and Wu, M. (2013). Surprisingly extensive mixed phylogenetic
and ecological signals among bacterial operational taxonomic units.
Nucleic Acids Research 41, 5175–5188. doi:10.1093/NAR/GKT241
Kunin, V., Engelbrektson, A., Ochman, H., and Hugenholtz, P. (2010).
Wrinkles in the rare biosphere: pyrosequencing errors can lead to
artificial inflation of diversity estimates. Environmental Microbiology
12, 118–123. doi:10.1111/J.1462-2920.2009.02051.X
Kwok, S., and Higuchi, R. (1989). Avoiding false positives with PCR.
Nature 339, 237–238. doi:10.1038/339237A0
Laramie, M. B., Pilliod, D. S., and Goldberg, C. S. (2015). Charecterising the
distribution of an endangered salmonid using eDNA analysis. Biological
Conservation 183, 29–37. doi:10.1016/J.BIOCON.2014.11.025
Lee, Z. M. P., Bussema, C., and Schmidt, T. M. (2009). Rrndb: documenting
the number of rRNA and tRNA genes in bacteria and archaea. Nucleic
Acids Research 37, D489–D493. doi:10.1093/NAR/GKN689
Lewis, G. D., and Metcalf, T. G. (1988). Polyethylene glycol precipitation
for recovery of pathogenic viruses, including hepatitis a virus and human
rotavirus, from oyster, water, and sediment samples. Applied and
Environmental Microbiology 54, 1983–1988.
Lim, Y. W., Kim, B. K., and Kim, C. (2010). Assessment of soil fungal
communities using pyrosequencing. Journal of Microbiology (Seoul,
Korea) 48, 284–289. doi:10.1007/S12275-010-9369-5
Logue, J. B., Bürgmann, H., and Robinson, C. T. (2008). Progress in the
ecological genetics and biodiversity of freshwater bacteria. Bioscience
58, 103–113. doi:10.1641/B580205
Lozupone, C. A., and Knight, R. (2005). UniFrac: a new phylogenetic method
for comparing microbial communities. Applied and Environmental
Microbiology 71, 8228–8235. doi:10.1128/AEM.71.12.8228-8235.2005
Lozupone, C. A., and Knight, R. (2008). Species divergence and the
measurement of microbial diversity. FEMS Microbiology Reviews 32,
557–578. doi:10.1111/J.1574-6976.2008.00111.X
Lozupone, C. A., Hamady, M., and Knight, R. (2006). UniFrac: an online
tool for comparing microbial community diversity in a phylogenetic
context. BMC Bioinformatics. 7, 371. doi:10.1186/1471-2105-7-371
Mahé, F., Rognes, T., Quince, C., de Vargas, C., and Dunthorn, M. (2014).
Swarm: robust and fast clustering method for amplicon-based studies.
PeerJ 2, e593. doi:10.7717/PEERJ.593
Mahmoudi, N., Slater, G. F., and Fulthorpe, R. R. (2011). Comparison of
commercial DNA extraction kits for isolation and purification of
bacterial and eukaryotic DNA from PAH-contaminated soils. Canadian
Journal of Microbiology 57, 623–628. doi:10.1139/W11-049
Mardis, E. R. (2008). Next generation DNA sequencing methods. Annual
Review of Genomics and Human Genetics 9, 387–402. doi:10.1146/
ANNUREV.GENOM.9.081307.164359
Martin-Laurent, F., Philipot, L., Hallet, S., Chaussod, R., Germon, J. C.,
Soulas, G., and Catroux, G. (2001). DNA extraction from soils: old bias
for new microbial diversity analysis methods. Applied and Environmental
Microbiology 67, 2354–2359. doi:10.1128/AEM.67.5.2354-2359.2001
A guide to eDNA-based biodiversity surveys
Medinger, R., Nolte, V., and Pandey, R. M. (2010). Diversity in a hidden
world: potential and limitation of next-generation sequencing for sur-
veys of molecular diversity of eukaryotic microorganisms. Molecular
Ecology 19, 32–40. doi:10.1111/J.1365-294X.2009.04478.X
Mende, D. R., Waller, A. S., Sunagawa, S., Järvelin, A. I., Chan, M. M.,
Arumugam, M., Raes, J., and Bork, P. (2012). Assessment of metage-
nomic assembly using simulated next generation sequencing data. PLoS
One 7, e31386. doi:10.1371/JOURNAL.PONE.0031386
Merkes, C., McCalla, S. G., Jensen, N. R., Gaikowski, M. P., and Amberg, J. J.
(2014). Persistence of DNA in carcasses, slime and avian feces may affect
interpretation of environmental DNA data. PLoS One 9, e113346.
doi:10.1371/JOURNAL.PONE.0113346
Meyer, F., Paarmann, D., D’Souza, M., Olson, R., Glass, E. M., Kubal, M.,
Paczian, T., Rodriguez, A., Stevens, R., Wilke, A., Wilkening, J., and
Edwards, R. A. (2008). The metagenomics RAST server – a public
resource for the automatic phylogenetic and functional analysis
of metagenomes. BMC Bioinformatics 9, 386. doi:10.1186/1471-
2105-9-386
Muyzer, G., de Waal, E. C., and Uitterlinden, A. G. (1993). Profiling of
complex microbial populations by denaturing gradient gel electro-
phoresis analysis of polymerase chain reaction-amplified genes coding
for 16S rRNA. Applied and Environmental Microbiology 59, 695–700.
Nilsson, R. H., Kristiansson, E., Ryberg, M., Hallenberg, N., and Larsson, K.
H. (2008). Intraspecific ITS variability in the kingdom Fungi as
expressed in the international sequence databases and its implications
for molecular species identification. Evolutionary Bioinformatics. 4,
193–201.
Patel, R. K., and Jain, M. (2012). NGS QC toolkit: a toolkit for quality
control of next generation sequencing data. PLoS One 7, e30619.
doi:10.1371/JOURNAL.PONE.0030619
Pennisi, E. (2011). Will computers crash genomics? Science 331, 666–668.
doi:10.1126/SCIENCE.331.6018.666
Petrosino, J. F., Highlander, S., Luna, R. A., Gibbs, R. A., and Versalovic, J.
(2009). Metagenomic pyrosequencing and microbial identification.
Clinical Chemistry 55, 856–866. doi:10.1373/CLINCHEM.2008.107565
Pilliod, D. S., Goldberg, C. S., Arkle, R. S., and Waits, L. P. (2013).
Estimating occupancy and abundance of stream amphibians using
environmental DNA from filtered water samples. Canadian Journal of
Fisheries and Aquatic Sciences 70, 1123–1130. doi:10.1139/CJFAS-
2013-0047
Polz, M. F., and Cavanaugh, C. M. (1998). Bias in template-to-product ratios
in multi-template PCR. Applied and Environmental Microbiology 64,
3724–3730.
Porazinska, D. L., Giblin-Davis, R. M., Faller, L., Farmerie, W., Kanzaki, N.,
Morris, K., Powers, T. O., Tucker, A. E., Sung, W., and Thomas, W. K.
(2009). Evaluating high-throughput sequencing as a method for metage-
nomic analysis of nematode diversity. Molecular Ecology Resources 9,
1439–1450. doi:10.1111/J.1755-0998.2009.02611.X
Resources Inventory Committee (1997). ‘Fish Collection Methods and
Standards.’ (RIC: Vancouver, BC.)
Riaz, T., Shehzad, W., Viari, A., Pompanon, F., Taberlet, P., and Coissac, E.
(2011). ecoPrimers: inference of new DNA barcode markers from whole
genome sequence analysis. Nucleic Acids Research 39, e145.
doi:10.1093/NAR/GKR732
Robertson, C. E., Harris, J. K., Wagner, B. D., Granger, D., Browne, K.,
Tatem, B., Feazel, L. M., Park, K., Pace, N. R., and Frank, D. N. (2013).
Explicet: graphical user interface software for metadata-driven manage-
ment, analysis and visualization of microbiome data. Bioinformatics 29,
3100–3101. doi:10.1093/BIOINFORMATICS/BTT526
Rooney, A. P., and Ward, T. J. (2005). Evolution of a large ribosomal RNA
multigene family in filamentous fungi: birth and death of a concerted
evolution paradigm. Proceedings of the National Academy of Sciences of
the United States of America 102, 5084–5089. doi:10.1073/PNAS.
0409689102
Marine and Freshwater Research M
Roossinck, M. J. (2012). Plant virus metagenomics: biodiversity and ecology.
Annual Review of Genetics 46, 359–369. doi:10.1146/ANNUREV-
GENET-110711-155600
Rozen, S., and Skaletsky, H. (1999). Primer3 on the WWW for general users
and biologist programmers. Bioinformatics methods and protocols 132,
365–386. doi:10.1385/1-59259-192-2:365
Salter, S. J., Cox, M. J., Turek, E. M., Calus, S. T., Cookson, W. O., Moffatt,
M. F., Turner, P., Parkhill, J., Loman, N. J., and Walker, A. W. (2014).
Reagent and laboratory contamination can critically impact sequence-
based microbiome analyses. BMC Biology 12, 87. doi:10.1186/S12915-
014-0087-Z
Schoch, C. L., Seifert, K. A., Huhndorf, S., Robert, V., Spouge, J. L.,
Levesque, C. A., and Chen, W.Fungal Barcoding Consortium (2012).
Nuclear ribosomal internal transcribed spacer (ITS) region as a universal
DNA barcode marker for Fungi. Proceedings of the National Academy of
Sciences of the United States of America 109, 6241–6246. doi:10.1073/
PNAS.1117018109
Shaw, J. L. A., Monis, P., Fabris, R., Ho, L., Braun, K., Drikas, M., and
Cooper, A. (2014). Assessing the impact of water treatment on bacterial
biofilms in drinking water distribution systems using high-throughput
DNA sequencing. Chemosphere 117, 185–192. doi:10.1016/J.CHEMO
SPHERE.2014.06.077
Shendure, J., and Ji, H. (2008). Next-generation DNA sequencing. Nature
Biotechnology 26, 1135–1145. doi:10.1038/NBT1486
Shokralla, S., Spall, J. L., Gibson, J. F., and Hajibabaei, M. (2012). Next-
generation sequencing technologies for environmental DNA research.
Molecular Ecology 21, 1794–1805. doi:10.1111/J.1365-294X.2012.
05538.X
Sigsgaard, E. E., Carl, H., Moller, P. R., and Thomsen, P. F. (2015).
Monitoring the near extinct European weather loach in Denmark based
on environmental DNA from water. Biological Conservation 183,
46–52. doi:10.1016/J.BIOCON.2014.11.023
Sogin, M. L., Morrison, H. G., Huber, J. A., Welch, D. M., Huse, S. M., Neal,
P. R., Arrieta, J. M., and Herndl, G. J. (2006). Microbial diversity in the
deep sea and the underexplored ‘rare biosphere’. Proceedings of the
National Academy of Sciences of the United States of America 103,
12115–12120. doi:10.1073/PNAS.0605127103
Spear, S. F., Groves, J. D., Williams, L. A., and Waits, L. P. (2015). Using
environmental DNA methods to improve detectability in a hellbender
(Cryptobranchus alleganiensis) monitoring program. Biological
Conservation 183, 38–45. doi:10.1016/J.BIOCON.2014.11.016
Stackebrandt, E., and Goebel, B. M. (1994). Taxonomic note: a place for
DNA–DNA reassociation and 16S rRNA sequence analysis in the
present species definition in bacteriology. International Journal of
Systematic Bacteriology 44, 846–849. doi:10.1099/00207713-44-4-846
Strayer, D. L., Eviner, V. T., Jeschke, J. M., and Pace, M. L. (2006).
Understanding the long-term effects of species invasions. Trends in
Ecology & Evolution 21, 645–651. doi:10.1016/J.TREE.2006.07.007
Strickler, K. M., Fremier, A. K., and Goldberg, C. S. (2015). Quantifying
effects of UV-B, temperature, and pH on eDNA degradation in aquatic
microcosms. Biological Conservation 183, 85–92. doi:10.1016/
J.BIOCON.2014.11.038
Sweeney, B. W., Battle, J. M., Jackson, J. K., and Dapkey, T. (2011). Can
DNA barcodes of stream macroinvertebrates improve descriptions of
community structure and water quality? Journal of the North American
Benthological Society. 30, 195–216. doi:10.1899/10-016.1
Taberlet, P., Coissac, E., Hajibabaei, M., and Rieseberg, L. H. (2012).
Environmental DNA. Molecular Ecology 21, 1789–1793. doi:10.1111/
J.1365-294X.2012.05542.X
Takahara, T., Minamoto, T., Yamanaka, H., Doi, H., and Kawabata, Z.
(2012). Estimation of fish biomass using environmental DNA. PLoS One
7, e35868. doi:10.1371/JOURNAL.PONE.0035868
Thomsen, P. F., Kielgast, J., Iversen, L. L., Moller, P. R., Rasmussen, M.,
and Willerslev, E. (2012a). Detection of a diverse marine fish fauna
N Marine and Freshwater Research J. A. Shaw et al.

using environmental DNA from seawater samples. PLoS One 7, e41732.
doi:10.1371/JOURNAL.PONE.0041732
Thomsen, P. F., Kielgast, J., Iversen, L. L., Wiuf, C., Rasmussen, M.,
Gilbert, T. P., Orlando, L., and Willerslev, E. (2012b). Monitoring
endangered freshwater biodiversity using environmental DNA. Molecular
Ecology 21, 2565–2573. doi:10.1111/J.1365-294X.2011.05418.X
Turner, C. R., Barnes, M. A., Xu, C. C. Y., Jones, S. E., Jerde, C. L., and
Lodge, D. M. (2014). Particle size distribution and optimal capture of
aqueous macrobial eDNA. Methods in Ecology and Evolution 5,
676–684. doi:10.1111/2041-210X.12206
Venter, J. C., Remington, K., Heidelberg, J. F., Halpern, A. L., Rusch, D.,
Eisen, J., Wu, D., Paulsen, I., Nelson, K. E., Nelson, W., Fouts, D. E.,
Levy, S., Knap, A. H., Lomas, M. W., Nealson, K., White, O., Peterson,
J., Hoffman, J., Parson, R., Baden-Tilson, H., Pfannkoch, C., Rogers, Y.,
and Smith, O. H. (2004). Environmental genome shotgun sequencing of
the Sargasso Sea. Science 304, 66–74. doi:10.1126/SCIENCE.1093857
von Atzigen, N., and Kennedy, S. (2011). Application note: determining the
best homogenization protocol for any soil. (Mo Bio Laboratories Inc.:
Carlsbad, CA.) Available at http://www.mobio.com/images/custom/
file/pdf/12855_apnote_web.pdf [Verified 1 October 2014].
Yoccoz, N. G., Brathen, K. A., Gielly, L., Haile, J., Edwards, M. E., Goslar,
T., Von Stedingk, H., Brysting, A. K., Coissac, E., Pompanon, F.,
Sonstebo, J. H., Miquel, C., Valentini, A., De Bello, F., Chave, J.,
Thuiller, W., Wincker, P., Cruaud, C., Gavory, F., Rasmussen, M.,
Gilbert, M. T. P., Orlando, L., Brochmann, C., Willerslev, E., and
Taberlet, P. (2012). DNA from soil mirrors plant taxonomic and growth
form diversity. Molecular Ecology 21, 3647–3655. doi:10.1111/J.1365-
294X.2012.05545.X
Yu, D. W., Ji, Y., Emerson, B. C., Wang, X., Ye, C., Yang, C., and Ding, Z.
(2012). Biodiversity soup: metabarcoding of arthropods for rapid
biodiversity assessment and biomonitoring. Methods in Ecology and
Evolution 3, 613–623. doi:10.1111/J.2041-210X.2012.00198.X

www.publish.csiro.au/journals/mfr

The author has requested enhancement of the downloaded file. All in-text references underlined in blue are linked to publications on ResearchGate.