Trends in Food Science & Technology 91 (2019) 294–304
Contents lists available at ScienceDirect
Trends in Food Science & Technology
journal homepage: www.elsevier.com/locate/tifs
Review
Determining the provenance and authenticity of seafood: A review of T
current methodologies
Karthik Gopia, Debashish Mazumdera,b,c,∗, Jesmond Sammuta,b, Neil Saintilanc
a Centre for Ecosystem Science, School of Biological, Earth and Environmental Sciences, The University of New South Wales, UNSW, Sydney, 2052, Australia
bAustralian Nuclear Science and Technology Organisation, Locked Bag 2001, Kirrawee DC, NSW, 2232, Australia
c Department of Environmental Sciences, Macquarie University, Sydney, NSW, Australia
A R T I C L E I N F O A B S T R A C T
Keywords: Background: Globally, food provenance has become a concern for government authorities, the seafood industry
Seafood and consumers due to increasing food safety and authenticity requirements. Wild-catch fisheries and aquaculture
Authentication are both important industries; aquaculture is seen as an opportunity to strengthen food security for the growing
Traceability global population. However, unregulated aquaculture can expose consumers to health risks from pathogens,
Aquaculture antibiotics and banned chemicals. Consumers and retailers, and the reputation of the global seafood industry, is
Fingerprinting
affected by food fraud through species substitution and the exchange of aquaculture produce with wild-caught
product and vice versa. To ensure consumer confidence and to allow authorities to effectively enforce regula-
tions and contain risks, methods to determine the species, production methods and geographic origin of seafood
need to be readily available.
Scope and approach: This review summarises the currently available and emerging methodologies to determine
the provenance and authenticity of seafood. The main focus of this review is to give an overview of the methods
that could potentially be used by authorities to enforce regulations and to contain risks, and for the seafood
industry to self-regulate and protect itself from food fraud.
Key findings and conclusions: The most common methods used are DNA profiling, fatty acid profiling, different
methods of inductively coupled plasma spectrometry and stable isotope analysis. Additionally, methods such as
blockchain, radio frequency identification and x-ray fluorescence through Itrax are currently being tested for
their effectiveness in determining seafood provenance. However, these methods have drawbacks and it is likely
that a combination of methods would be best suited to determine the provenance of seafood considering its
complex supply chain.
1. Introduction Aquaculture production has been steadily increasing since 1950 and
it currently produces more seafood than from capture fisheries (FAO,
Food provenance is emerging as a major concern globally for con- 2018). If managed correctly, aquaculture can provide an important
sumers, the seafood industries and regulatory bodies due to food safety source of protein for the increasing global population (Naylor et al.,
issues, fraudulent relabelling of products (Ulrich et al., 2015), and a 2000). Seafood is a key component of a nutritional diet because of its
desire, by consumers, to know the origin and production methods essential macro- and micro-nutrients, including omega-3 fatty acids and
(Kelly, Heaton, & Hoogewerff, 2005). Additionally, consumers are in- vitamins (Sioen, Matthys, De Backer, Van Camp, & Henauw, 2007;
creasingly aware of the impact of food production on the environment World Health Organization, 2003). The seafood market generates an
and have been faced with disease outbreaks (e.g. Avian flu and white estimated $151 billion USD per annum, demonstrating the high global
spot disease in prawns) along with malpractice from producers demand for seafood (Organisation for Economic Cooperation and
(Galimberti et al., 2013). These issues have compounded as countries Development, 2017). As seafood consumption increases and wild fish-
increase food imports to focus on more profitable sectors, whilst pro- eries deplete, the market will be driven towards aquaculture-based
viding consumers access to exotic and otherwise seasonal goods all year seafood products (Botsford, Castilla, & Peterson, 1997; Kearney, 2010).
round (Marianela, Dieter, Michael, Wolfgang, & Wolfgang, 2013). Consequently, the price of wild-caught seafood might increase as the
∗ Corresponding author. Centre for Ecosystem Science, School of Biological, Earth and Environmental Sciences, The University of New South Wales, UNSW, Sydney,
2052, Australia.
E-mail address: debashish.mazumder@ansto.gov.au (D. Mazumder).
https://doi.org/10.1016/j.tifs.2019.07.010
Received 3 May 2019; Received in revised form 24 June 2019; Accepted 22 July 2019
Available online 24 July 2019
0924-2244/ Crown Copyright © 2019 Published by Elsevier Ltd. All rights reserved.
K. Gopi, et al. Trends in Food Science & Technology 91 (2019) 294–304
supply depletes. This increase in price has led exploitative importers
and exporters to intentionally mislabel food to profiteer by charging
more for a cheaper product thus harming the reputation of legitimate
businesses. Over recent years, several countries have reported this
practice of substituting expensive seafood products with cheaper com-
modities leading to what can be described as seafood fraud (Buck,
2007; Ottavian et al., 2012). Countries, including Taiwan and Germany,
have detected and protected against seafood fraud using methods such
as identification of species through DNA profiling (Hsieh et al., 2010;
Huang et al., 2014; Rehbein, 2008), although this approach does not
determine production method. Methods of determining food prove-
nance are required to combat food fraud, which is estimated to cost the
global food industry between $30 to $40 billion USD per year
(PricewaterhouseCoopers, 2016).
Apart from affecting the profitability of the seafood industry seafood
fraud also raises concerns regarding the safety, hygiene and authenti- Fig. 1. Bar graph showing the number of keystone studies examined in this
city of fraudulently imported seafood (Furness & Osman, 2006; Ulrich review for each method from 1990 to 2019. All of the papers in this graph
et al., 2015). Human health has been placed at risk by the presence of examined the use of methods to determine the provenance of seafood.
pathogens and other banned substances such as antibiotics found in
imported seafood (Feldhusen, 2000). More recently there have been studied for their use in determining seafood provenance; and
increased concerns regarding microplastics which are being ingested by 3) To provide practical recommendations regarding the appropriate
marine species consumed by humans (Choy & Drazen, 2013; de Sá, application of methods.
Oliveira, Ribeiro, Rocha, & Futter, 2018; Rochman et al., 2016;
Rochman et al., 2015; Van Cauwenberghe & Janssen, 2014). However,
it is not yet clear what effects microplastics have on human health 2. Methodology
(Rochman et al., 2016). Seafood fraud can also irreparably damage the
reputation of products and their producers. For instance, if a human A literature search was conducted using Boolean search terms on
health issue is caused by a fraudulent product presented as a genuine ‘Elsevier’, ‘Science Direct’, ‘ProQuest’ and ‘Scopus’ to find papers from
one (e.g. Atlantic salmon (Salmo salar) being substituted with rainbow the last 20 years of research that were relevant to the literature review.
trout (Oncorhynchus mykiss)), it can lead to a recall of the genuine The Boolean search terms included the name of the methodology (in-
product. This can have a drastic effect on legitimate businesses that rely cluding abbreviations) along with the term ‘traceability OR provenance
on exporting or importing seafood. These issues, along with the afore- OR fingerprint OR authentication’. This was done to ensure that the
mentioned risks to human health, highlight the need for an accurate search captured a wide range of papers that may have used different
system of tracing seafood back to its origin. Contaminated seafood can keywords. From this search, “keystone” papers that indicate the efficacy
be effectively recalled if authorities can track the origin of the offending of a method were examined, and their contents were analysed (Fig. 1).
product. Furthermore, consumers have demonstrated their preference The terms ‘traceability OR provenance OR fingerprint OR authentica-
for high quality and environmentally friendly seafood (Kelly et al., tion’ were chosen as they were commonly used interchangeably in the
2005). Recently, there has been a public outcry over the links between literature. Generally, the keywords ‘traceability’, ‘fraud’, ‘safety’, and
seafood production and slavery along with labour rights abuse ‘authentication’ appeared together with some papers using ‘fingerprint’
(Kittinger et al., 2017). These issues, in addition to the fraudulent or ‘provenance’ in place of ‘authentication’ and ‘traceability’.
mislabelling of seafood, highlight the importance of seafood prove-
nance. The complex supply chain needed for the product to reach the 3. Results and discussion
consumer compounds the myriad of issues faced by the seafood in-
dustry (Leal, Pimentel, Ricardo, Rosa, & Calado, 2015). Therefore, ac- 3.1. DNA profiling
curate methods to determine seafood origin and production methods
are necessary to ensure consumer confidence in imported seafood. DNA profiling has been used extensively by several producers and
There are several methods that are currently available for de- authorities to determine the provenance of food (Scarano & Rao, 2014),
termining the provenance of seafood, along with a few emerging due to the many advantages provided by DNA profiling. Using DNA
methodologies. Previously there have been several reviews that focused profiling to identify animal species in food is ideal because the beha-
on highlighting the applicability and potential shortcomings of several viour of DNA is not species-dependent and is predictable, food samples
of these methods (Leal et al., 2015; Primrose, Woolfe, & Rollinson, which have been heated up to 120 °C in the cooking process can still be
2010; Verrez-Bagnis, 2017). However, there are only a limited number analysed, and the diversity of DNA allows for the differentiation of
of studies exploring emerging technologies in food provenance and closely-related species and even subspecies (Lenstra, 2003). De-
fewer still exploring their role in determining the provenance of seafood termining the species of plants or animals used in food products is an
(Badia-Melis, Mishra, & Ruiz-García, 2015; Schröder, 2008). Therefore, important part of detecting food adulteration as well as allowing for the
a systematic review which covers the advantages and disadvantages of detection of food substitution (Hsieh et al., 2010; Huang et al., 2014;
the current methods is necessary in order to compare them with Rehbein, 2008).
emerging technologies which are being tested for their use in de- Generally, DNA profiling is based on two types of markers; hy-
termining seafood provenance. In this review, we highlight some im- bridisation-based markers and Polymerase Chain Reaction (PCR)-based
portant limitations and constraints, both technical and commercial, and markers. Out of these two methods, PCR is considered faster and more
identify instances along the food supply chain where seafood fraud can accurate when compared to other methods (Labra et al., 2004;
occur. Our aims are, therefore: Teletchea, Maudet, & Hänni, 2005). There is a variety of other methods
also used for DNA profiling in food traceability; however, the lack of
1) To provide an updated overview of the methods available to de- standardisation and universality of methods is an issue (Galimberti
termine seafood provenance; et al., 2013). This is especially important as new species are introduced
2) Highlight new and emerging technologies that are currently being into the market as demonstrated by Rehbein (2008). When these
295
K. Gopi, et al. Trends in Food Science & Technology 91 (2019) 294–304
species are introduced to new markets it is likely that they will also be wild-caught S. salar from Ireland. The farmed samples were genetically
subject to seafood fraud. For example, high-value species such as ar- different as most of the fingerlings for the farming stock were likely
rowtooth flounder (Atheresthes stomias) and yellowfin sole (Limanda from Norway or Scotland, while the wild-caught samples were geneti-
aspera) were being substituted with cheaper species known as pangasius cally adapted to Ireland. The exact cause of genetic variation between
(including Pangasius hypophthalmus and Pangasius bocourti) which were the same species has not been researched exhaustively and some studies
recently introduced to the European market. Rehbein (2008) focused on argue that the variation is caused by adapting to local conditions while
profiling these cheaper species to detect seafood fraud and noted that others argue that it is caused by genetic drift (McGinnity et al., 1997).
without a centralised public database it is difficult to use DNA profiling Although DNA markers may vary according to different geographic
as a method for detecting food substitution. However, this is currently locations, the causes of genetic variation between the same species need
being addressed with databases such as the Barcode of Life Data System to be explored to determine whether or not DNA profiling can be used
(BOLD) and the Fish Barcode of Life (FISH-BOL) (Barbuto et al., 2010). as a method for determining the production method of seafood.
DNA profiling has been used to investigate several instances of food
substitution in the seafood market (Barbuto et al., 2010; Carrera et al., 3.2. Fatty acid profiling
2000; Cross & Challanain, 1991; Russell et al., 2000). Barbuto et al.
(2010) collected 45 samples of “palombo” (Mustelus mustelus or Mus- Fatty acids are aliphatic monocarboxylic acids which are contained
telus asterias) from the Italian market and characterised their DNA in, or derived from, an animal or vegetable fat, oil or wax. Natural fatty
profile using PCR. When the profiles were compared to samples in acids can be saturated or unsaturated and contain 4 to 28 carbons (Nic,
BOLD, it was found that 35 out of the 45 samples were substituted with Hovorka, Jirat, Kosata, & Znamenacek, 2014). Fatty acid profiling ty-
a different species. Only three of the collected samples could be related pically uses various methods of chromatography or spectroscopy to
directly to “palombo” (they were all M. mustelus). While this study determine the fatty acid composition of samples (James et al., 2011).
demonstrates that DNA profiling can determine the species of a product However, in seafood provenance research, it is typically bundled with
to help prevent seafood fraud, it also highlights the lack of standardised Stable Isotope Analysis (SIA) (Busetto et al., 2008; Zhang, Liu, Li, &
databases containing the DNA profile of different species of seafood. Zhao, 2017). While this method is slower than some of the others due to
Thus, the DNA profile of the majority of seafood species need to be the time taken to prepare samples for analysis (Budge, Iverson, Bowen,
analysed and curated. Nevertheless, the method is able to show that a & Ackman, 2002), it has shown promise in discriminating between
seafood product has been mislabelled or substituted. production methods and geographic origin of samples (Bergström,
Similarly, Carrera et al. (2000) used PCR to identify and compare 1989; Grahl-Nielsen, Jacobsen, Christophersen, & Magnesen, 2010;
the DNA profile of Atlantic salmon (S. salar) and rainbow trout (O. Grigorakis, Alexis, Taylor, & Hole, 2002; Nemova, Fokina, Nefedova,
mykiss). The main purpose of this study was to determine if DNA pro- Ruokolainen, & Bakhmet, 2013; Olsen, Grahl-Nielsen, & Schander,
filing could differentiate between these two species because S. salar is 2009; Ricardo et al., 2015b).
often substituted with O. mykiss, as it is a cheaper product. The study The dietary lipids of penaeid shrimp are reflected in the fatty acid
showed that DNA profiling could cost-effectively discriminate between composition of samples. This was demonstrated by Lim, Ako, Brown,
these two species. The study also determined that the DNA profile of and Hahn (1997) who examined the differences in the fatty acid com-
smoked samples was similar to the raw samples for both species, de- position of Penaeus vannamei fed seven different diets. While the aim of
monstrating the robustness of DNA. Additionally, Russell et al. (2000) the study was not directly related to seafood provenance, it provides a
used DNA profiling to differentiate between ten different species of basis for discriminating between shellfish. The diets which contained
salmon from a number of different locations. This study focused on high levels of unsaturated fatty acids caused the fatty acid profile of the
using reactive enzymes to trim the DNA down to specific locations that P. vannamei to differ. This suggests that the differences in diets between
differentiated between the species. Russell et al. (2000) were able to farms and wild-caught shellfish should vary significantly, allowing for
differentiate the species using a DNA fragment that can be amplified them to be distinguished. Budge et al. (2002) performed a large-scale
even in processed samples. This shows that the current process of DNA study which analysed the fatty acids of 28 different species of fish and
profiling can be further improved while reducing the cost of analysis invertebrates from three geographic locations. Using fatty acid profiling
and improving throughput. Hsieh et al. (2010) were able to replicate they were able to determine the provenance of 16 of these species with
this process for four species of puffer fish. The method was less ex- an accuracy greater than 98%. While they were able to distinguish
pensive than a direct sequencing analysis and was able to produce re- between geographic origin using fatty acid profiles influenced by diets,
sults within 9 h. The result suggests that this methodology is ideal for the variation within-species was not as high as that of among-species
use in seafood identification, especially when a rapid analysis is re- variation, suggesting that differentiating between production methods
quired for regulatory bodies to make decisions. Huang et al. (2014) may be difficult. This paper showcases the utility of fatty acid profiling
further expanded on this methodology by determining the gene markers in distinguishing between different geographic locations of seafood.
for 12 different species of puffer fish and adding them to the GenBank Grigorakis et al. (2002) analysed the fatty acids on gilthead sea bream
database. The authors remarked that the short gene marker regions (Sparus aurata) using a VARIAN 3300 gas chromatograph. While this
were remarkably stable against environmental stress and can be used study was not specifically aimed at determining the provenance of S.
with PCR. This is especially important once the species are processed aurata, it still showed that there was a significant difference in the lipid
and used as part of other products. These studies show the benefits of and fatty acid profile of wild-caught samples compared to farmed
utilising PCR and characterising certain sections of DNA, the cost and samples. In all cases, it was found that the farmed samples had higher
benefit analysis of each method is important as food fraud already costs lipid content than their wild counterparts. Additionally, it is important
the food industry a large amount of money. Therefore, methods which to note that the fatty acid profile of the farmed samples reflected the
can be used to detect and prevent food fraud need to be accurate while composition of their feed. This shows that fatty acid profiling can also
being affordable. be suitable for distinguishing between farmed and wild-caught fish.
In addition to differentiating between species, it is also important to However, Grigorakis et al. (2002) noted that there was a significant
differentiate between the same species from different production seasonal variation in wild-caught S. aurata samples. This could become
methods. This is especially important for species like Lates calcarifer as a potential issue if the samples’ fatty acid profile matches that of one
they are cultured and wild-caught in multiple countries (e.g. Australia, from an entirely different region due to these seasonal variations. This
Malaysia, Indonesia, Taiwan, etc) and sold under different names in is a well-known issue and was already highlighted by Bergström (1989).
certain markets (e.g. Asian seabass vs Barramundi) (FAO, 2006). Cross The study by Bergström (1989) showed that the seasonal changes in
and Challanain (1991) successfully discriminated between farmed and fatty acid composition and total lipid content of wild and farmed S.
296
K. Gopi, et al. Trends in Food Science & Technology 91 (2019) 294–304
salar vary significantly before smolting. Smolting is a process by which 3.3. Elemental profiling
the salmon undergoes morphological changes before migrating to the
sea (McCormick, Hansen, Quinn, & Saunders, 1998). The wild-caught S. Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) is a high-
salar had a significantly higher concentration of saturated fats than the precision method used to detect the concentrations of trace elements in
farmed S. salar. The study also found that some polyunsaturated fatty geological samples (Jenner, Longerich, Jackson, & Fryer, 1990; Reid,
acids were higher in wild-caught samples when compared to farmed Horn, Longerich, Forsythe, & Jenner, 1999). ICP-MS has been utilised in
samples. Additionally, significant changes were found in the total lipid the field of food traceability and can distinguish between the geo-
content of both farmed and wild-caught fish between seasons. While graphic origins of seafood (Dunphy, Millet, & Jeffs, 2011; Ricardo et al.,
Bergström (1989) did not aim to discriminate between the farmed and 2015a; Sorte, Etter, Spackman, Boyle, & Hannigan, 2013). This method
wild-caught salmon, the results are still significant for food traceability. requires samples to be digested before being analysed and certain ele-
The results from this study suggest that there are significant seasonal ments, such as mercury, may require further preparation (Hight &
changes to the lipid content of fish and that this needs to be considered Cheng, 2006; Ricardo et al., 2015a). Inductively Coupled Plasma-
when attempting to trace the geographic origin and production method Atomic Emission Spectroscopy (ICP-AES) also known as Inductively
of seafood using fatty acid profiling. Nemova et al. (2013) were able to Coupled Plasma-Optical Emission Spectrometry (ICP-OES) is similar to
show that the fatty acid profile of the gill lipids in blue mussels (Mytilus ICP-MS and has also been utilised in seafood provenance studies (K. A.
edulis) varied greatly with salinity. This variation in the fatty acid Anderson, Hobbie, & Smith, 2010; Li, Boyd, Odom, & Dong, 2013;
composition can be used to distinguish between different production Smith & Watts, 2009). These methods can be performed on both soft
methods such as intertidal zone collection or aquaculture. The results tissue and hard calcified tissue; the methodology and interpretation of
from these studies suggest that fatty acid profiling can be a valuable results are different according to the type of tissue sample being ex-
tool to trace the production method of seafood. amined (K. A. Anderson et al., 2010; Dunphy et al., 2011).
Ricardo et al. (2015b) used the fatty acid profile of cockles (Cer- Sorte et al. (2013) used laser-ablation inductively-coupled mass
astoderma edule) to determine if it could distinguish between eight spectrometry (LA-ICP-MS) to determine the geographical origin of blue
different sites which were from a single coastal lagoon in Portugal. The mussel (Mytilus edulis). The study was able to distinguish between five
study found that fatty acid profiling was able to discriminate between different sites, roughly 50 km apart, with greater than 50% accuracy in
collection sites within the same coastal region. However, it was unable all cases. The accuracy was higher for juvenile samples and when close
to discriminate between the same production areas. It is important to sites were grouped together it increased the accuracy to 97%. Similarly,
note that these sites are defined by the amount of Escherichia coli pre- Dunphy et al. (2011) used LA-ICP-MS to assign juvenile mussels to their
sent in the flesh of the C. edule. Therefore, discriminating between these collection sites around northern New Zealand with over 60% accuracy
collection and production areas can be an important marketing tool for in all cases. Seven elemental ratios were proven to be the most reliable
the C. edule industry as it will be able to prove that the product is of when used with discriminant analysis. This study is important as it
higher value due to a lower amount of contamination. Additionally, demonstrates that not all elements need to be determined when dis-
Ricardo et al. (2015b) note that the method can be even more cost tinguishing the provenance of seafood samples. Ricardo et al. (2015a)
effective by focusing on a particular part of the cockles. Olsen et al. used ICP-MS to determine the concentration of aluminium, barium,
(2009) used fatty acid profiling to differentiate clams (Astarte sulcata) calcium, cadmium, copper, magnesium, manganese, lead, strontium
between four areas, grouped according to proximity, in Norway. The and zinc in cockle (Cerastoderma edule) shells. These 10 elements were
study found that fatty acid profiling is advantageous when comparing analysed using three different statistical analyses and it was found that
closely-related sites as it gives high-resolution results, allowing it to these elements were capable of distinguishing samples as little as 1 km
detect small changes in populations that are geographically close. apart. Cubadda, Raggi, and Coni (2006) used ICP-MS to determine the
However, it is important to note that the composition of the fatty acids elemental concentration of 16 different elements present in Medi-
present in the samples is what discriminates between sites and not the terranean mussel (Mytilus galloprovincialis) and a few other species of
presence or absence of certain fatty acids. Similarly, Grahl-Nielsen et al. fin-fish. Using PCA they were able to clearly distinguish between three
(2010) used fatty acid profiling to discriminate between five locations farming sites in close proximity. Similarly, Costas-Rodríguez, Lavilla,
along the coastal regions of Norway. The data collected in the study and Bendicho (2010) determined 40 elements in 158 samples and used
were analysed using Partial Least Square (PLS) and Principal Compo- it to distinguish between five collection sites of M. galloprovincialis. The
nent Analysis (PCA) to distinguish between groups. The study found study used linear discriminant analysis (LDA) and found that reducing
that the five sites had distinct fatty acid compositions, further sup- the number of elements down to 16 reduced the accuracy down to
porting the conclusion drawn by Olsen et al. (2009). Hence, it is clear 95.6% but had a number of incorrect predictions.
that fatty acid profiling can play a distinct role in distinguishing sam- ICP-AES has also been utilised as a method for determining the
ples from geographical locations. elemental composition of samples to differentiate between geographic
Fatty acid profiling has also been used in conjunction with SIA to locations. K. A. Anderson et al. (2010) used a combination of ICP-AES
distinguish between the geographic origins of seafood. A study con- and SIA to differentiate between farmed and wild-caught king salmon
ducted by Busetto et al. (2008), comparing the fatty acid composition of (Oncorhynchus tshawytscha), coho salmon (Oncorhynchus kisutch) and
wild and farmed turbots, found that the monounsaturated and poly- Atlantic salmon (Salmo salar). Using PCA and Canonical Discriminant
unsaturated fatty acids of the farmed fish were higher than in wild- Analysis (CDA) it was possible to discriminate between the production
caught fish. This would enable a measurable difference in the fatty acid methods of each species using only 12 elements. This demonstrates the
composition of the two production methods. Both the fatty acid and SIA efficacy of the ICP-AES methodology in determining the provenance of
results were used in a PCA to discriminate between geographic loca- seafood. Similarly, Li et al. (2013) utilised ICP-AES to discriminate
tions and the production methods of turbots. Similarly, Zhang et al. between three different locations of channel catfish (Ictalurus punctatus)
(2017) used fatty acid profiling and SIA of carbon and nitrogen to and blue catfish (Ictalurus furcatus) using 11 elements. While the study
differentiate between sea cucumbers (Apostichopus japonicus) collected was able to successfully discriminate the geographic origin of the
from seven sites in northern China. While the stable carbon and ni- samples, it was noted that the elemental profile could vary or even be
trogen values were capable of distinguishing between five out of the similar across various locations with similar elemental compositions.
seven sites, it had an overlap between two, likely due to the fish having Additionally, as the elemental composition of water will vary within the
similar food sources. However, by using both the stable isotope values same geographic origin, it is important to have a large database of the
and fatty acid profiles, the PCA was able to distinguish between the elemental composition of a species for a given country, before ele-
seven sites clearly. mental profiling can be used for seafood provenance (Li et al., 2013). A
297
K. Gopi, et al. Trends in Food Science & Technology 91 (2019) 294–304
similar issue was raised by Smith and Watts (2009) when they under- the geographic origin of samples found that utilising SIA alone was not
took a large scale study to determine the provenance of prawns (mostly always accurate (Carter et al., 2015; Turchini et al., 2009). Further
P. vannamei and Penaeus monodon) from 8 countries. The samples were studies are needed to investigate the application of SIA combined with
collected from around ten regions in each country and analysed using other analytical techniques to assess if the approach can determine the
ICP-MS. The overall accuracy of the database was around 90% and the geographic region of samples accurately. Kim, Kumar, Hwang, et al.
study calls for databases to be built for different geographic origins to (2015) used SIA to differentiate between the geographic origin of
provide a comparative base to allow for accurate prawn provenance shrimp and hairtail fish. The study found that the Korean hairtail fish
determination (Smith & Watts, 2009). (family Trichiuridae) and their international counterparts had sig-
nificantly different isotopic signatures for both δ13C and δ15N. This
3.4. Stable isotope analysis suggests that SIA can be used to discriminate between the same species
from various geographical locations. Gamboa-Delgado et al. (2014)
SIA uses isotopes to distinguish between samples. Isotopes are ele- used SIA to differentiate between wild-caught and farmed Pacific white
ments which have the same number of protons but different numbers of shrimp (Litopenaeus vannamei) where there was more variability in the
neutrons. Stable isotopes are assimilated into animal tissues as they δ13C values of the wild-caught and farmed shrimp than in their δ15N
move up through the trophic chain (Fry, 1991; Mazumder, Wen, values. The variability of the δ15N values was significantly lower in the
Johansen, Kobayashi, & Saintilan, 2016; Peterson & Fry, 1987; Post, farmed samples than in the wild-caught samples. In general, the sam-
2002). Typically, the technique determines the stable isotopes of carbon ples which were wild caught were isotopically enhanced for both δ13C
(δ13C) and nitrogen (δ15N) but hydrogen and oxygen can also be de- and δ15N. Using both the δ13C and δ15N isotope values to discriminate
tected. Primary producers have a distinct isotopic signature and when between the samples is ideal as it had an accuracy of 99% in this study.
they are consumed by animals this signature assimilates into the tissues SIA is not without fault as noted by a study conducted to dis-
of the consumers through a process known as fractionation, which is criminate between cultured and wild caught turbot (Psetta maxima).
caused by the changes in the heavy to the light isotopic ratio The SIA results showed overlap between the carbon and nitrogen ratios
(Ehleringer, Rundel, & Nagy, 1986). The stable carbon isotopes indicate of wild-caught Danish turbots and cultured Spanish turbots (Busetto
the sources of nutrients, while the nitrogen isotopes indicate the trophic et al., 2008). This argument is supported by a number of studies such as
level of an organism in the food web. Because the isotopic values of a the review conducted by Primrose et al. (2010), which suggests that SIA
consumer are related to the composition of the diet (Fry, 2006; Kling, should be combined with trace element measurements to provide a
Fry, & O'Brien, 1992), the differences in such diets, stemming from comprehensive method for seafood traceability. Ortea and Gallardo
changes in farming practices or environmental conditions, would be (2015) found that combining the carbon and nitrogen stable isotope
reflected in the isotopic profile of consumer's muscle. ratios along with other elements such as arsenic and lead is the most
SIA for food provenance has shown positive results when dis- reliable method of tracing the production method and geographic
criminating between production methods (Carter, Tinggi, Yang, & Fry, origin respectively. Carter et al. (2015) used SIA to analyse both the
2015; Gamboa-Delgado et al., 2014; Kim, Kumar, & Shin, 2015; Ortea & water recovered from prawn samples as well as the prawns in order to
Gallardo, 2015; Turchini, Quinn, Jones, Palmeri, & Gooley, 2009). distinguish between Australian and imported prawns. The water had no
Molkentin, Meisel, Lehmann, and Rehbein (2007) utilised SIA of the correlation with their respective tissue samples and using the results
δ13C and δ15N ratio to differentiate between conventionally farmed, obtained from the water was not sufficient to distinguish between
wild caught and organically farmed Atlantic salmon. However, SIA Australian and imported samples; the isotopic analysis revealed sig-
alone was not able to distinguish between the three and the studies nificant differences in hydrogen and carbon isotopes. Turchini et al.
found that a combination of linoleic acid and SIA was necessary to (2009) used chemical analyses as well as carbon, nitrogen and oxygen
distinguish between the three different production methods of the stable isotopes to discriminate between different farms. This study
Atlantic salmon. Molkentin, Lehmann, Ostermeyer, and Rehbein (2015) showed the effectiveness of SIA in differentiating between farms in the
used stable isotopes and fatty acid analysis to distinguish between the same region. However, a study conducted on wild-caught and farmed
organic and conventionally farmed Atlantic salmon. This study showed Brazilian freshwater cachara suggested that the C/N ratio of these fish
that it was possible to distinguish between the production methods of varies seasonally, especially in the wet season. However, the C/N ratio
Atlantic salmon using SIA alone. Similarly, Fasolato, et al. (2010) used was indistinguishable in the dry season, suggesting that the farmed and
δ13C analysis of fat-free muscle to successfully distinguish between wild-caught samples would be indistinguishable during this time
farmed and wild-caught L. calcarifer, as the δ13C shows the feeding period. The lack of variation between the production methods needs to
habit of the fish. Fasolato et al. (2010) selected δ13C based on the study be studied to determine the cause (Sant’Ana, Ducatti, & Ramires, 2010).
by Sweeting, Barry, Polunin, and Jennings (2007) who concluded that Gopi et al. (2019b) used both SIA and elemental profiling, through X-
δ13C has minimal seasonal variation, especially in large predatory fish. ray fluorescence, and found that a combined model built using both
A recent study by Gopi et al. (2019b), showed that SIA alone was able to datasets had an accuracy of 81% and had no incorrect predictions out of
determine the provenance of L. calcarifer with over 84% accuracy and six. This result was repeated by Gopi et al. (2019a) for black tiger
only had two incorrect predictions out of six. This suggests that SIA, like prawns (P. monodon) where the combined dataset resulted in an accu-
many of the other methods, can distinguish the provenance of seafood. racy of> 98%. Both these studies showed that SIA can be utilised in
However, Gopi et al. (2019b) found a significant enrichment in the δ15N conjunction with other analytical techniques to provide an accurate
values of wild-caught L. calcarifer. This finding conflicts with the find- tool to authenticate the provenance of seafood. We believe that the
ings of Serrano, Blanes, and Orero (2007) as they found no significant higher accuracy shown in both Gopi et al. (2019b) and Gopi et al.
enrichment of δ15N values for wild-caught gilthead sea bream (Sparus (2019a) is due to the way in which they analysed the data; both studies
aurata). The difference in enrichment could be due to interspecies used multiple statistical analyses designed to distinguish between
variability, as the studies analysed data from two different species. groups to determine the provenance of their samples, leading to an
However, the findings of Gopi et al. (2019b) agree with Moreno-Rojas, increased overall accuracy. While these results are promising, the
Tulli, Messina, Tibaldi, and Guillou (2008) who analysed rainbow trout methodology needs to be studied further as they found variations in the
(O. mykiss) using SIA to discriminate between samples fed with fishmeal results depending on the species.
and plant-based protein diets. The study found that the samples fed the
fishmeal were significantly enriched in δ15N due to the fish-based
proteins present in their diet.
On the other hand, other studies which have used SIA to determine
298
K. Gopi, et al. Trends in Food Science & Technology 91 (2019) 294–304
4. New and emerging technologies and methodologies multiple sources. Therefore, to ensure that all partners are complying
with the correct procedures necessary for blockchain, a method is re-
4.1. Blockchain quired for testing the actual provenance of seafood. This can act as a
deterrent for food fraud and provide additional information to con-
The blockchain technology was exposed to the world through the sumers so that they can be assured that they are purchasing a properly
bitcoin cryptocurrency in 2008 (Bhardwaj & Kaushik, 2018). Block- labelled product.
chain technology involves a combination of principles from the 1960s
based on timestamping digital documents using crypto signatures 4.2. Radio frequency identification
(Haber & Stornetta, 1990), a decentralised storage system where re-
cords cannot be purged (R. Anderson, 1996) and encrypting files to Radio frequency identification (RFID) is also an emerging prove-
prevent access from untrusted/unauthorised machines (Schneier & nance technology. RFID stores information, such as serial numbers and
Kelsey, 1998). Hence, blockchain has the ability to operate without a place of origin, on a microchip attached to an antenna. The information
centralised and trusted authority to authenticate transactions or records emitted through the radio waves is scanned and converted to digital
and prevents them from being erased. Additionally, the system is made information which can be displayed on the scanner (Doukidis,
secure from unauthorised parties through the use of mathematical Pramatari, & Kelepouris, 2007). The advantage of utilising RFID is that
problems, which needs a substantial amount of computational power to it can automatically capture data needed for traceability with minimal
solve (Galvez, Mejuto, & Simal-Gandara, 2018). changes to the business process, leading to reduced expenditure needed
The advantage of utilising blockchain over traditional bookkeeping to implement the system (Doukidis et al., 2007). RFID does provide
methods is the ability to encrypt end-to-end traceability and allow the certain advantages over traditional barcodes such as being automatic,
consumer to access this information easily via the internet (Galvez having a long lifespan, being robust and being able to store up to 32
et al., 2018). When this technology is applied to the food supply chain it kilobytes of data in each tag (Costa et al., 2013). RFID also offers some
will allow for the storage of a wide range of data, from GPS coordinates advantages such as being hidden out of sight as the tags do not need to
of where fish were caught to the batch number of a fish produced be in sight of the reader to work, allowing for them to be stored inside
through aquaculture. Any data that are deemed to be important can be the product or containers (Costa et al., 2013). Additionally, in the
added into the blockchain system by the members of the business majority of environments RFID can be read successfully in the first scan
network. Once all members of the network authenticate the entered in 99.5%–100% of cases (Texas Instruments, 2006). In Europe, this
data they cannot be altered, allowing for a permanent record of all data system has been implemented by several seafood producers as part of
on that particular food product (Galvez et al., 2018). Some businesses, their compliance with European regulations on traceability (Schröder,
like Walmart in the USA, are already in the process of implementing 2008). Furthermore, the technology has also been used to monitor the
blockchain into their produce (Yiannas, 2018). When the methodology temperature fluctuations inside a refrigerated vehicle transporting
was applied to their mango supply chain, the time taken to trace the frozen tilapia fillets (Tingman, Jian, & Xiaoshuan, 2010). This would
end product back to the farm reduced from almost seven days using allow retailers to predict the shelf-life of the product based on tem-
traditional methods down to 2.2 s using blockchain (Yiannas, 2018). perature fluctuations (Tingman et al., 2010). As food traceability re-
The blockchain cannot only keep an immutable copy of records but can quires all partners in a supply chain to work cooperatively, it might be
also make accessing those records much more efficient. expensive for small and medium-sized enterprises to implement a
Implementing blockchain into the seafood supply chain can be a system on par with the bigger enterprises. RFID can play a significant
relatively simple and cost-effective process because of cloud-based role in cases like these as a potential traceability system which uses
models (Korpela, Hallikas, & Dahlberg, 2017). As an example, a wild- RFID might only require a scanner and a computer (Doukidis et al.,
catch supplier of skipjack tuna (Katsuwonus pelamis) would enter the 2007). Therefore, RFID is a potentially cost-effective solution which can
catchment area of a batch of fish which is then transported to the be compete with methods such as blockchain. However, this does not
processor. The primary processor would then fillet the fish, ensuring prevent seafood fraud as RFID chips with fraudulent information can be
that this process is entered into the blockchain before shipping the created easily. Therefore, methods which can authenticate the pro-
produce to a cannery. The blockchain of final product (i.e. canned tuna) duction methods and geographic origin of seafood are necessary to
would contain all this information, which can then be easily accessed ensure that products are correctly labelled using RFID.
by regulatory bodies and industry partners. An added benefit of this
technique of bookkeeping, is that this information can be passed along 4.3. X-ray fluorescence through Itrax
to the consumer, allowing them to make informed decisions. The main
advantage of using blockchain for seafood provenance is that it is X-ray fluorescence (XRF) through Itrax is a method that provides the
tamper proof. To falsify a record entered into a blockchain system, you elemental composition of samples and is typically utilised for sediment
would need control of over 51% of the nodes in the supply chain (Tian, core scanning or dendrochronology (Gunnarson, Linderholm, &
2018). When a new transaction or record is added into the system it is Moberg, 2011; Keegan et al., 2008; Zuo, 2013). XRF through Itrax has
verified by the nodes in the whole system, before it is added to the the potential to be a cost-effective method of determining the elemental
blockchain (Tian, 2018). However, for the blockchain system to work in profile of seafood samples. While it may not give quantified values of
a real-world context, it needs to be taken up by every single organisa- elements, it can detect the presence of up to 31 different elements and
tion along the supply chain. If it is not implemented in this manner it should be readily available in geology labs.
would make it difficult for the chain of custody of a seafood product to Recently, Gadd, et al. (2018) developed a method that determines
remain tamperproof. the elemental composition of organic soft-tissue samples. This method
As mentioned previously, the seafood supply chain is incredibly was used to developed a seafood provenance model for Asian seabass
complex (Leal et al., 2015). The initial product can go through multiple (L. calcarifer) (Gopi et al., 2019b). This study managed to correctly
processors and supply chain actors before it reaches the consumer, and classify both the geographical origin and production method of L. cal-
along the way some information can become lost. For instance, a sec- carifer samples with no incorrect predictions using both SIA and ele-
ondary producer might obtain the same species of fish from multiple mental data gathered using Itrax. A similar approach was taken for tiger
sources and fillet them before sending it to the distributor. If a box of prawns (P. monodon) to distinguish the provenance of the samples with
fillets is almost full then it is likely that it will be filled with fillets from high certainty (Gopi et al., 2019a). However, while these results are
other sources. This information might not be stored in the blockchain as promising, it is important to note that the datasets used to train the
fillets are indistinguishable from each other even if they are from models in these two papers were limited. The differences between the
299
K. Gopi, et al. Trends in Food Science & Technology 91 (2019) 294–304
from similar locations more effectively than by SIA. However, the
seasonal variability in the fatty acid composition of different species
needs to be explored. If this is not done, then there can be potential
overlaps between geographical locations and production methods.
Additionally, while there are variations to the fatty acid compositions
according to both production methods and geographic locations, it
could potentially be manipulated through controlling the diets. Fur-
thermore, to determine the actual origin of unknown samples, a data-
base containing the fatty acid profiles of different species is needed.
Once these issues have been addressed, fatty acid profiling can play an
important role in seafood provenance.
ICP-MS or ICP-AES has been shown to be an important tool for
seafood provenance (K. A. Anderson et al., 2010; Sorte et al., 2013).
While ICP-MS can determine the concentration of most elements, some
studies attempted to reduce the number of elements needed to clearly
distinguish between locations. This is a key issue that needs to be ad-
dressed, as ICP-MS requires samples to be prepared before analysis and
cannot measure all elements in one sample run. Therefore, determining
a large number of elements in each sample will require a significant
amount of time. For ICP-MS to become a standard methodology for
discriminating between the geographical locations and production
methods of seafood, the elements necessary to discriminate the prove-
nance of each species needs to be determined. Additionally, a database
of the elemental profile of a species from a geographic region needs to
be recorded to allow for comparison of samples to determine food
fraud. Once these issues are addressed, elemental profiling can play a
Fig. 2. Decision tree to choose a method suitable for determining seafood role in fingerprinting the geographic locations of seafood.
provenance and authenticating species for different scenarios. SIA is differentiated by being able to distinguish not only the geo-
graphic origin of seafood but also the production method with a rela-
different geographic origins and production methods need to be ex- tively high accuracy (Gopi et al., 2019b; Molkentin et al., 2007).
plored further. Once the causes of differences are determined the However, most of the studies had to use another method to distinguish
knowledge can be used to adjust and train the model further to reduce between samples that had overlaps in their stable carbon or nitrogen
incorrect predictions. Currently it is unclear whether or not the inter- isotopic ratios (Busetto et al., 2008; Ortea & Gallardo, 2015). These
pretation of the results will change according to the type of tissues overlaps may occur due to farmed species being fed a similar diet, wild-
being scanned because the studies have only used muscle tissue. caught species migrating across borders and anthropogenic pollution.
Therefore, this method should be tested on more species and different Additionally, seasonal variability in the stable isotopes may cause these
types of tissues in order to determine its utility as a tool for seafood overlaps to occur. As the isotopic composition of a product is influenced
provenance. by factors such as feed and environmental conditions, it is possible that
the isotopic composition can be manipulated by controlling these fac-
4.4. The current state of seafood provenance tors. While this type of fraud is unlikely to occur due to the costs in-
volved, it would make detecting these instances of fraud difficult using
Currently, regulatory bodies and authorities have several methods SIA alone. Therefore, from our experience and from the literature (Gopi
available to them to determine the provenance of seafood. New meth- et al., 2019b, 2019a; Primrose et al., 2010), it is suggested that SIA is
odologies are constantly being developed and can often supplement the combined with other methods when used for seafood provenance as it
current suite of tools. Therefore, to determine the ideal method for helps to distinguish between samples with the same isotopic signature.
different scenarios the decision tree we developed can be consulted Evolving technologies such as blockchain and RFID are being tested
(Fig. 2). Furthermore, to understand the advantages and shortcomings across various food sectors, including seafood, and have the potential to
of each method a table has been provided summarising the findings of complement the other analytical techniques described in this review.
this review (Table 1). While these methods can provide rapid access to records of a product to
DNA profiling has clear advantages when it comes to detecting food the end user, there are still a few limitations which need to be ad-
substitution as the species of even heavily processed samples can be dressed. While there have been a few implementations of these methods
determined. While the issue of universality is being tackled by data- into small scale supply chains (Schröder, 2008), they remain relatively
bases such as BOLD and FISH-BOL (Barbuto et al., 2010), the issue of untested in the more complex supply chains which exist in the case of
standardisation still exists. Standardisation is a much more difficult some seafood products. For instance, some seafood processors may use
issue to tackle as it requires all labs to follow a standard procedure. products from several countries; however, if only a few producers have
Even with these drawbacks, DNA profiling remains one of the most implemented blockchain or RFID then then authenticity of the final
recommended methods of analyses when it comes to detecting food processed fillets may be difficult to confirm. Additionally, because
fraud, at least for determining the species. However, the inability of blockchain and RFID are essentially bookkeeping methods, provenance
DNA profiling to distinguish between farmed and wild-caught samples determination methods need to be available to ensure that all partners
of the same species, that are from the same location, makes it difficult in a supply chain are held responsible for the input of data. The XRF
to use when farmed samples are passed off as wild-caught. through Itrax to use the elemental composition of the samples to de-
Fatty acid profiling has been used to distinguish the provenance of a termine the provenance of samples holds promise (Gopi et al., 2019a;
wide variety of seafood. It is not only useful for distinguishing between 2019b). While it may not provide the quantified values of elements, for
production methods but also to discriminate between geographical the purpose of determining provenance alone this is not a disadvantage.
origins (Busetto et al., 2008; Nemova et al., 2013). The high-resolution However, the reasons for variations between geographic locations and
output provided by fatty acid profiling can distinguish between seafood production methods need to be determined before the method can be
300
K. Gopi, et al. Trends in Food Science & Technology 91 (2019) 294–304
301
Table 1
A summary of the advantages and disadvantages of each methodology examined in this review.
Method Advantage Drawbacks Turnaround time Non-destructive
sampling?
DNA profiling • determine species • cannot be used to distinguish between production methods Relatively rapid No• distinguish between geographic locations • incomplete DNA profiles available for cross referencing• DNA profiling is offered by many labs • different labs use different techniques• can be used on processed seafood• relatively cost-effective
Fatty acid profiling • can be used to distinguish between production methods and • samples require extensive preparation Relatively rapid No
geographic origins • seasonal changes in fatty acid composition unclear at the moment• relatively cost-effective
ICP – MS/AES/OES • can be used to distinguish between production methods and • each element requires different preparation Relatively rapid No
geographic locations • the elements necessary to distinguish between production methods and geographic• relatively cost-effective depending on the number of locations differ according to species
elements • elemental composition can potentially be manipulated to have the same composition
as a genuine product by fraudulent producers
• interpretation of results can vary according to type of tissue being analysed
SIA • can be used to determine both production methods and • certain geographic locations and production methods may overlap with each other Slower than other methods No
geographic locations • isotopic composition can potentially be manipulated through feed and environmental• relatively cost-effective conditions
Blockchain • can be used to determine species, production methods and • still untested in complex supply chains like seafood Almost instantaneous Yes
geographic locations
• relatively easy to implement
RFID • can store species, production and location information • difficult to implement in processed fillets of fish Almost instantaneous Yes• has been utilised previously with seafood and has been
shown as a cost-effective method
XRF through Itrax • can distinguish between production methods and geographic • does not provide quantitative values of elements like ICP-MS/AES/OES Relatively rapid Yes
locations • elemental composition can potentially be manipulated to have the same composition• relatively cost-effective for processing large batches of as a genuine product by fraudulent producers
samples
K. Gopi, et al. Trends in Food Science & Technology 91 (2019) 294–304
Fig. 4. Another example of seafood fraud occurring, where a fraudulent seller
Fig. 3. An example of seafood fraud occurring, where the primary processor has has substituted local wild-caught Atlantic salmon with imported rainbow trout
substituted local wild-caught seafood with farmed and imported seafood. and Atlantic salmon.
widely recommended for the purpose of seafood provenance determi- standard for detecting food fraud. Fatty acid profiling has proved useful
nation. in discriminating between the production methods of seafood as well as
Determining the provenance of seafood becomes difficult as seafood their geographic locations. The only major drawback of this method is
fraud becomes increasingly complex. There are several opportunities the seasonal variability in the fatty acid composition of seafood.
along the seafood supply chain for fraudulent businesses to substitute Without a thorough understanding of how this might affect the de-
products. For instance, during primary processing there is an opportu- termination of provenance, it is difficult to recommend the use of fatty
nity to add products from multiple origins and different species into a acid profiling as a standalone method. Elemental profiling using ICP-MS
single batch (Fig. 3). This will increase profit because imported pro- and ICP-AES has shown promise when it comes to determining the
ducts are often cheaper than local those sourced locally. In order to provenance of seafood. To further improve the method, the elements
detect this type of food fraud a combination of methods should be used. that are vital in discriminating between geographic locations and pro-
For instance, by using SIA and XRF through Itrax, the geographic origin duction methods need to be determined to reduce the processing time
and production method of the fillets can be determined (Gopi et al., and improve throughput. Stable isotope analysis of carbon and nitrogen
2019b). Utilising multiple methods is recommended here due to the has proven effective in distinguishing the provenance of seafood when
limitations of the current methods. Similarly, a species substitution is combined with other methodologies. In addition, new technologies and
often carried out by fraudulent businesses in order to increase profit methodologies are being developed for the purpose of seafood prove-
margins (Fig. 4). DNA profiling is the only current method that can nance. Without additional tests, it is difficult to comment on the utility
detect species substitution; to determine origin a combination of cur- of these methods on a large scale to combat widespread problems. As
rent methods should be used. These are just a few examples of how mentioned previously, the supply chain for seafood is highly complex.
seafood fraud can occur along the supply chain. However, tackling We believe that a combination of different methodologies is ideal for
these issues require multiple methods moment and addressing some of seafood provenance. Using a combination of techniques has advantages
the shortfalls of the current methods will allow for better accuracy in such as being able to predict the source of origin with a higher degree of
determining the provenance of seafood. accuracy than a single methodology alone. Additionally, a specific
model that is developed for determining the provenance of seafood will
5. Conclusion add to this and provide regulatory bodies with the tools and techniques
necessary for seafood traceability. This can also be used for compliance
There are a number of methodologies available to regulatory bodies testing of methods like blockchain and RFID to ensure that the details
and the seafood industry to determine the provenance of seafood with stored in the system and being relayed to consumers are accurate.
varying degrees of accuracy. Additionally, new technologies that enable Overall, this is a pivotal time for seafood provenance research, as
the consumer to track the source of their seafood are being developed several methodologies are currently in development. The cost-effec-
and tested. DNA profiling can be utilised to detect food substitution as tiveness of these methodologies needs to be determined before they
well as, to a lesser extent, to discriminate between geographic locations become a widespread and common method for seafood provenance. By
of seafood. However, without a complete database of DNA pro les of determining the provenance of seafood, food fraud can be detected andfi
different species, it can be difficult to recommend this method for de- prevented, strengthening market and customer confidence.
termining the provenance of seafood, although it remains the de-facto Additionally, it will protect consumer health as offending products
302
K. Gopi, et al. Trends in Food Science & Technology 91 (2019) 294–304
which cause outbreaks of disease can be recalled more quickly. Cubadda, F., Raggi, A., & Coni, E. (2006). Element fingerprinting of marine organisms by
Furthermore, regulatory bodies can be assured that imported seafood is dynamic reaction cell inductively coupled plasma mass spectrometry. Analytical and
Bioanalytical Chemistry, 384, 887–896.
from certified sources allowing for imported seafood to meet the needs Doukidis, G., Pramatari, K., & Kelepouris, T. (2007). RFID‐enabled traceability in the food
of consumers. Overall, food provenance is more important now than supply chain. Industrial Management & Data Systems, 107, 183–200.
ever as food producers gear up to feed a growing global population. Dunphy, B. J., Millet, M. A., & Jeffs, A. G. (2011). Elemental signatures in the shells of
early juvenile green-lipped mussels (Perna canaliculus) and their potential use for
larval tracking. Aquaculture, 311, 187–192.
Declaration of interest Ehleringer, J. R., Rundel, P. W., & Nagy, K. A. (1986). Stable isotopes in physiological
ecology and food web research. Trends in Ecology & Evolution, 1, 42–45.
The authors of this paper do not have any nancial or personal FAO (2006). Fishery and aquaculture statistics. Food and agriculture organizaion of thefi
united nations (pp. 77). .
relationships with other people or organisations that could in- FAO (2018). The state of World Fisheries and Aqauculture 2018 - meeting the sustainable
appropriately influence this review. The authors are not employed by development goals. Food and agriculture organization of the united nations (pp. 3–4). .
the seafood industry and do not have any financial interest in the in- Fasolato, L., Novelli, E., Salmaso, L., Corain, L., Camin, F., Perini, M., et al. (2010).
Application of nonparametric multivariate analyses to the authentication of wild and
dustry. farmed European sea bass (Dicentrarchus labrax). Results of a survey on fish sampled
in the retail trade. Journal of Agricultural and Food Chemistry, 58, 10979–10988.
Funding Feldhusen, F. (2000). The role of seafood in bacterialfoodborne diseases. Microbes and
Infection, 2, 1651–1660.
Fry, B. (1991). Stable isotope diagrams of freshwater food webs. Ecology, 72, 2293–2297.
This work was supported by the Australian Institute of Nuclear Fry, B. (2006). Using stable isotope tracers. Stable isotope ecology. Vol. 521. Stable isotope
Science and Engineering – Residential Student Scholarship. ecology (pp. 40–75). Springer.
Furness, A., & Osman, K. A. (2006). 1 - developing traceability systems across the food
supply chain: An overview. Improving traceability in food processing and distribution
Acknowledgements (pp. 3–25). Woodhead Publishing.
Gadd, P., Gopi, K., Sammut, J., Saintilan, N., Crawford, J., & Mazumder, D. (2018). Itrax
The authors would like to thank Prof Marie-Claude, Prof Hendrik micro X-ray fluorescence (μXRF) for soft biological tissues. Methods (Orlando), 5,1267–1271.
Heijnis and Dr Suzanne Hollins from the Australian Nuclear Science and Galimberti, A., De Mattia, F., Losa, A., Bruni, I., Federici, S., Casiraghi, M., et al. (2013).
Technology Organisation (ANSTO) for their continued support for DNA barcoding as a new tool for food traceability. Food Research International, 50,
seafood provenance research at ANSTO and the research partnership 55–63.
Galvez, J. F., Mejuto, J. C., & Simal-Gandara, J. (2018). Future challenges on the use of
with the University of New South Wales, Macquarie University, blockchain for food traceability analysis. TRAC Trends in Analytical Chemistry, 107,
National Measurement Institute, and Sydney Fish Market. 222–232.
Gamboa-Delgado, J., Molina-Poveda, C., Godínez-Siordia, D. E., Villarreal-Cavazos, D.,
Ricque-Marie, D., & Cruz-Suárez, L. E. (2014). Application of stable isotope analysis
References to differentiate shrimp extracted by industrial fishing or produced through aqua-
culture practices. Canadian Journal of Fisheries and Aquatic Sciences, 71, 1520–1528.
Anderson, R. (1996). The eternity service. Proceedings of PRAGOCRYPT. Vol. 96. Gopi, K., Mazumder, D., Sammut, J., Saintilan, N., Crawford, J., & Gadd, P. (2019a).
Proceedings of PRAGOCRYPT (pp. 242–252). Combined use of stable isotope analysis and elemental profiling to determine pro-
Anderson, K. A., Hobbie, K. A., & Smith, B. W. (2010). Chemical pro ling with modeling venance of black tiger prawns (Penaeus monodon). Food Control, 95, 242–248.fi
di erentiates wild and farm-raised salmon. Journal of Agricultural and Food Chemistry, Gopi, K., Mazumder, D., Sammut, J., Saintilan, N., Crawford, J., & Gadd, P. (2019b).ff
58, 11768–11774. Isotopic and elemental profiling to trace the geographic origins of farmed and wild-
Badia-Melis, R., Mishra, P., & Ruiz-García, L. (2015). Food traceability: New trends and caught Asian seabass (Lates calcarifer). Aquaculture, 502, 56–62.
recent advances. A review. Food Control, 57, 393–401. Grahl-Nielsen, O., Jacobsen, A., Christophersen, G., & Magnesen, T. (2010). Fatty acid
Barbuto, M., Galimberti, A., Ferri, E., Labra, M., Malandra, R., Galli, P., et al. (2010). DNA composition in adductor muscle of juvenile scallops (Pecten maximus) from five
barcoding reveals fraudulent substitutions in shark seafood products: The Italian case Norwegian populations reared in the same environment. Biochemical Systematics and
of “palombo”(Mustelus spp.). Food Research International, 43, 376–381. Ecology, 38, 478–488.
Bergström, E. (1989). E ect of natural and arti cial diets on seasonal changes in fatty Grigorakis, K., Alexis, M. N., Taylor, K. D. A., & Hole, M. (2002). Comparison of wild andff fi
acid composition and total body lipid content of wild and hatchery-reared Atlantic cultured gilthead sea bream (Sparus aurata); composition, appearance and seasonal
salmon (Salmo salar L.) parr-smolt. Aquaculture, 82, 205–217. variations. International Journal of Food Science and Technology, 37, 477–484.
Bhardwaj, S., & Kaushik, M. (2018). Blockchain—technology to drive the future. Smart Gunnarson, B. E., Linderholm, H. W., & Moberg, A. (2011). Improving a tree-ring re-
computing and informatics (pp. 263–271). Springer. construction from west-central scandinavia: 900 years of warm-season temperatures.
Botsford, L. W., Castilla, J. C., & Peterson, C. H. (1997). The management of sheries and Climate Dynamics, 36, 97–108.fi
marine ecosystems. Science, 277, 509–515. Haber, S., & Stornetta, W. S. (1990). How to time-stamp a digital document. Conference on
Buck, E. H. (2007). Seafood marketing: Combating fraud and deception. Congressional the theory and application of cryptography (pp. 437–455). Springer.
research service. Library of Congress. Hight, S. C., & Cheng, J. (2006). Determination of methylmercury and estimation of total
Budge, S. M., Iverson, S. J., Bowen, W. D., & Ackman, R. G. (2002). Among-and within- mercury in seafood using high performance liquid chromatography (HPLC) and in-
species variability in fatty acid signatures of marine fish and invertebrates on the ductively coupled plasma-mass spectrometry (ICP-MS): Method development and
Scotian Shelf, Georges Bank, and southern Gulf of St. Lawrence. Canadian Journal of validation. Analytica Chimica Acta, 567, 160–172.
Fisheries and Aquatic Sciences, 59, 886–898. Hsieh, C.-H., Chang, W.-T., Chang, H. C., Hsieh, H.-S., Chung, Y.-L., & Hwang, D.-F.
Busetto, M. L., Moretti, V. M., Caprino, F., Giani, I., Bellagamba, F., Moreno-Rojas, J. M., (2010). Puffer fish-based commercial fraud identification in a segment of cytochrome
et al. (2008). Authentication of farmed and wild turbot (Psetta maxima) by fatty acid b region by PCR–RFLP analysis. Food Chemistry, 121, 1305–1311.
and isotopic analyses combined with chemometrics. Journal of Agricultural and Food Huang, Y.-R., Yin, M.-C., Hsieh, Y.-L., Yeh, Y.-H., Yang, Y.-C., Chung, Y.-L., et al. (2014).
Chemistry, 56, 2742–2750. Authentication of consumer fraud in Taiwanese fish products by molecular trace
Carrera, E., García, T., Céspedes, A., González, I., Fernández, A., Asensio, L. M., et al. evidence and forensically informative nucleotide sequencing. Food Research
(2000). Identification of smoked Atlantic salmon (Salmo salar) and rainbow trout International, 55, 294–302.
(Oncorhynchus mykiss) using PCR-restriction fragment length polymorphism of the James, G. O., Hocart, C. H., Hillier, W., Chen, H., Kordbacheh, F., Price, G. D., et al.
p53 gene. Journal of AOAC International, 83, 341–346. (2011). Fatty acid profiling of Chlamydomonas reinhardtii under nitrogen deprivation.
Carter, J. F., Tinggi, U., Yang, X., & Fry, B. (2015). Stable isotope and trace metal com- Bioresource Technology, 102, 3343–3351.
positions of Australian prawns as a guide to authenticity and wholesomeness. Food Jenner, G. A., Longerich, H. P., Jackson, S. E., & Fryer, B. J. (1990). ICP-MS — a powerful
Chemistry, 170, 241–248. tool for high-precision trace-element analysis in Earth sciences: Evidence from ana-
Choy, C. A., & Drazen, J. C. (2013). Plastic for dinner? Observations of frequent debris lysis of selected U.S.G.S. reference samples. Chemical Geology, 83, 133–148.
ingestion by pelagic predatory shes from the central north paci c. Marine Ecology Kearney, J. (2010). Food consumption trends and drivers. Philosophical Transactions of thefi fi
Progress Series, 485, 155–163. Royal Society B: Biological Sciences, 365, 2793.
Costa, C., Antonucci, F., Pallottino, F., Aguzzi, J., Sarriá, D., & Menesatti, P. (2013). A Keegan, E., Richter, S., Kelly, I., Wong, H., Gadd, P., Kuehn, H., et al. (2008). The pro-
review on agri-food supply chain traceability by means of RFID technology. Food and venance of Australian uranium ore concentrates by elemental and isotopic analysis.
Bioprocess Technology, 6, 353–366. Applied Geochemistry, 23, 765–777.
Costas-Rodríguez, M., Lavilla, I., & Bendicho, C. (2010). Classi cation of cultivated Kelly, S., Heaton, K., & Hoogewerff, J. (2005). Tracing the geographical origin of food:fi
mussels from Galicia (Northwest Spain) with European Protected Designation of The application of multi-element and multi-isotope analysis. Trends in Food Science &
Origin using trace element ngerprint and chemometric analysis. Analytica Chimica Technology, 16, 555–567.fi
Acta, 664, 121–128. Kim, H., Kumar, K. S., Hwang, S. Y., Kang, B.-C., Moon, H.-B., & Shin, K.-H. (2015). Utility
Cross, T. F., & Challanain, D. N. (1991). Genetic characterisation of Atlantic salmon of stable isotope and cytochrome oxidase I gene sequencing analyses in inferring
(Salmo salar) lines farmed in Ireland. Aquaculture, 98, 209–216. origin and authentication of hairtail fish and shrimp. Journal of Agricultural and Food
Chemistry, 63, 5548–5556.
303
K. Gopi, et al. Trends in Food Science & Technology 91 (2019) 294–304
Kim, H., Kumar, K. S., & Shin, K.-H. (2015). Applicability of stable C and N isotope by identification of species. Journal für Verbraucherschutz und Lebensmittelsicherheit, 3,
analysis in inferring the geographical origin and authentication of commercial fish 49–53.
(Mackerel, Yellow Croaker and Pollock). Food Chemistry, 172, 523–527. Reid, J. E., Horn, I., Longerich, H. P., Forsythe, L., & Jenner, G. A. (1999). Determination
Kittinger, J. N., Teh, L. C. L., Allison, E. H., Bennett, N. J., Crowder, L. B., Finkbeiner, E. of Zr and Hf in a flux-free fusion of whole rock samples using laser ablation in-
M., et al. (2017). Committing to socially responsible seafood. Science, 356, 912–913. ductively coupled plasma-mass spectrometry (LA-ICP-MS) with isotope dilution ca-
Kling, G. W., Fry, B., & O'Brien, W. J. (1992). Stable isotopes and planktonic trophic libration. Geostandards Newsletter, 23, 149–155.
structure in arctic lakes. Ecology, 73, 561–566. Ricardo, F., Génio, L., Costa Leal, M., Albuquerque, R., Queiroga, H., Rosa, R., et al.
Korpela, K., Hallikas, J., & Dahlberg, T. (2017). Digital supply chain transformation to- (2015a). Trace element fingerprinting of cockle (Cerastoderma edule) shells can reveal
ward blockchain integration. Proceedings of the 50th Hawaii international conference on harvesting location in adjacent areas. Scientific Reports, 5, 11932.
system sciences. Ricardo, F., Pimentel, T., Moreira, A. S. P., Rey, F., Coimbra, M. A., Rosário Domingues,
Labra, M., Miele, M., Ledda, B., Grassi, F., Mazzei, M., & Sala, F. (2004). Morphological M., et al. (2015b). Potential use of fatty acid profiles of the adductor muscle of
characterization, essential oil composition and DNA genotyping of Ocimum basilicum cockles (Cerastoderma edule) for traceability of collection site. Scientific Reports, 5,
L. cultivars. Plant Science, 167, 725–731. 11125.
Leal, M. C., Pimentel, T., Ricardo, F., Rosa, R., & Calado, R. (2015). Seafood traceability: Rochman, C. M., Browne, M. A., Underwood, A. J., Van Franeker, J. A., Thompson, R. C.,
Current needs, available tools, and biotechnological challenges for origin certifica- & Amaral-Zettler, L. A. (2016). The ecological impacts of marine debris: Unraveling
tion. Trends in Biotechnology, 33, 331–336. the demonstrated evidence from what is perceived. Ecology, 97, 302–312.
Lenstra, J. A. (2003). 2 - DNA methods for identifying plant and animal species in food. In Rochman, C. M., Tahir, A., Williams, S. L., Baxa, D. V., Lam, R., Miller, J. T., et al. (2015).
M. Lees (Ed.). Food authenticity and traceability (pp. 34–53). Woodhead Publishing. Anthropogenic debris in seafood: Plastic debris and fibers from textiles in fish and
Li, L., Boyd, C. E., Odom, J., & Dong, S. (2013). Identification of ictalurid catfish fillets to bivalves sold for human consumption. Scientific Reports, 5.
rearing location using elemental profiling. Journal of the World Aquaculture Society, Russell, V. J., Hold, G. L., Pryde, S. E., Rehbein, H., Quinteiro, J., Rey-Mendez, M., et al.
44, 405–414. (2000). Use of restriction fragment length polymorphism to distinguish between
Lim, C., Ako, H., Brown, C. L., & Hahn, K. (1997). Growth response and fatty acid salmon species. Journal of Agricultural and Food Chemistry, 48, 2184–2188.
composition of juvenile Penaeus vannamei fed different sources of dietary lipid. Sant'Ana, L. S., Ducatti, C., & Ramires, D. G. (2010). Seasonal variations in chemical
Aquaculture, 151, 143–153. composition and stable isotopes of farmed and wild Brazilian freshwater fish. Food
Marianela, F., Dieter, G., Michael, K., Wolfgang, L., & Wolfgang, C. (2013). Spatial de- Chemistry, 122, 74–77.
coupling of agricultural production and consumption: Quantifying dependences of de Sá, L. C., Oliveira, M., Ribeiro, F., Rocha, T. L., & Futter, M. N. (2018). Studies of the
countries on food imports due to domestic land and water constraints. Environmental effects of microplastics on aquatic organisms: What do we know and where should we
Research Letters, 8, 014046. focus our efforts in the future? The Science of the Total Environment, 645, 1029–1039.
Mazumder, D., Wen, L., Johansen, M. P., Kobayashi, T., & Saintilan, N. (2016). Inherent Scarano, D., & Rao, R. (2014). DNA markers for food products authentication. Diversity.
variation in carbon and nitrogen isotopic assimilation in the freshwater macro-in- Vol. 6.
vertebrate Cherax destructor. Marine and Freshwater Research, 67, 1928–1937. Schneier, B., & Kelsey, J. (1998). Cryptographic support for secure logs on untrusted
McCormick, S. D., Hansen, L. P., Quinn, T. P., & Saunders, R. L. (1998). Movement, machines. USENIX security symposium. Vol. 98. USENIX security symposium (pp.
migration, and smolting of Atlantic salmon (Salmo salar). Canadian Journal of Fisheries 53–62).
and Aquatic Sciences, 55, 77–92. Schröder, U. (2008). Challenges in the traceability of seafood. Journal für
McGinnity, P., Stone, C., Taggart, J. B., Cooke, D., Cotter, D., Hynes, R., et al. (1997). Verbraucherschutz und Lebensmittelsicherheit, 3, 45–48.
Genetic impact of escaped farmed atlantic salmon (Salmo salar L.) on native popu- Serrano, R., Blanes, M., & Orero, L. (2007). Stable isotope determination in wild and
lations: Use of DNA profiling to assess freshwater performance of wild, farmed, and farmed gilthead sea bream (Sparus aurata) tissues from the western Mediterranean.
hybrid progeny in a natural river environment. ICES Journal of Marine Science, 54, Chemosphere, 69, 1075–1080.
998–1008. Sioen, I., Matthys, C., De Backer, G., Van Camp, J., & Henauw, S. D. (2007). Importance of
Molkentin, J., Lehmann, I., Ostermeyer, U., & Rehbein, H. (2015). Traceability of organic seafood as nutrient source in the diet of Belgian adolescents. Journal of Human
fish – authenticating the production origin of salmonids by chemical and isotopic Nutrition and Dietetics, 20, 580–589.
analyses. Food Control, 53, 55–66. Smith, R. G., & Watts, C. A. (2009). Determination of the country of origin of farm-raised
Molkentin, J., Meisel, H., Lehmann, I., & Rehbein, H. (2007). Identification of organically shrimp (Family Penaeide) using trace metal profiling and multivariate statistics.
farmed Atlantic salmon by analysis of stable isotopes and fatty acids. European Food Journal of Agricultural and Food Chemistry, 57, 8244–8249.
Research and Technology = Zeitschrift für Lebensmittel-Untersuchung und -Forschung. A, Sorte, C. J. B., Etter, R. J., Spackman, R., Boyle, E. E., & Hannigan, R. E. (2013). Elemental
224, 535–543. fingerprinting of mussel shells to predict population sources and redistribution po-
Moreno-Rojas, J. M., Tulli, F., Messina, M., Tibaldi, E., & Guillou, C. (2008). Stable iso- tential in the Gulf of Maine. PLoS One, 8, e80868.
tope ratio analysis as a tool to discriminate between rainbow trout (O. mykiss) fed Sweeting, C. J., Barry, J. T., Polunin, N. V. C., & Jennings, S. (2007). Effects of body size
diets based on plant or fish-meal proteins. Rapid Communications in Mass Spectrometry, and environment on diet-tissue δ13C fractionation in fishes. Journal of Experimental
22, 3706–3710. Marine Biology and Ecology, 352, 165–176.
Naylor, R. L., Goldburg, R. J., Primavera, J. H., Kautsky, N., Beveridge, M. C. M., Clay, J., Teletchea, F., Maudet, C., & Hänni, C. J. T. i. b. (2005). Food and forensic molecular
et al. (2000). Effect of aquaculture on world fish supplies. Nature, 405, 1017–1024. identification: Update and challenges, Vol. 23, 359–366.
Nemova, N. N., Fokina, N. N., Nefedova, Z. A., Ruokolainen, T. R., & Bakhmet, I. N. Texas Instruments (2006). Wedge transponder RI-TRP-R9BK RI-TRP-W9WK. Reference
(2013). Modifications of gill lipid composition in littoral and cultured blue mussels Guide SCBU037. 25.
Mytilus edulis L. under the influence of ambient salinity. Polar Record, 49, 272–277. Tian, F. (2018). An information system for food safety monitoring in supply shains based on
Nic, M., Hovorka, L., Jirat, J., Kosata, B., & Znamenacek, J. (2014). IUPAC compendium of HACCP, blockchain and Internet of Things. WU Vienna University of Economics and
chemical terminology (2nd ed.). . The Gold Book, XML on-line corrected version: Business.
http://goldbook.iupac.org International Union of Pure and Applied Chemistry . Tingman, W., Jian, Z., & Xiaoshuan, Z. (2010). Fish product quality evaluation based on
Olsen, B. R., Grahl-Nielsen, O., & Schander, C. (2009). Population study of Astarte sulcata, temperature monitoring in cold chain. African Journal of Biotechnology, 9, 6146–6151.
da Costa, 1778, (Mollusca, Bivalvia) from two Norwegian fjords based on the fatty Turchini, G. M., Quinn, G. P., Jones, P. L., Palmeri, G., & Gooley, G. (2009). Traceability
acid composition of the adductor muscle. Biochemical Systematics and Ecology, 37, and discrimination among differently farmed fish: A case study on Australian murray
662–669. cod. Journal of Agricultural and Food Chemistry, 57, 274.
Organisation for Economic Cooperation and Development (2017). Fish and seafood. Ulrich, R. M., John, D. E., Barton, G. W., Hendrick, G. S., Fries, D. P., & Paul, J. H. (2015).
OECD-FAO agricultural outlook 2017-2026. Paris: OECD Publishing. A handheld sensor assay for the identification of grouper as a safeguard against
Ortea, I., & Gallardo, J. M. (2015). Investigation of production method, geographical seafood mislabeling fraud. Food Control, 53, 81–90.
origin and species authentication in commercially relevant shrimps using stable Van Cauwenberghe, L., & Janssen, C. R. (2014). Microplastics in bivalves cultured for
isotope ratio and/or multi-element analyses combined with chemometrics: An ex- human consumption. Environmental Pollution, 193, 65–70.
ploratory analysis. Food Chemistry, 170, 145–153. Verrez-Bagnis, V. (2017). Advances in authentication methods for seafood species iden-
Ottavian, M., Facco, P., Fasolato, L., Novelli, E., Mirisola, M., Perini, M., et al. (2012). Use tification in food products. In D. Montet, & R. C. Ray (Eds.). Food traceability and
of near-infrared spectroscopy for fast fraud detection in seafood: Application to the authenticity: Analytical techniques (pp. 196–215). New York: CRC Press.
authentication of wild european sea bass (Dicentrarchus labrax). Journal of Agricultural World Health Organization (2003). Food based dietary guidelines in the WHO European
and Food Chemistry, 60, 639–648. Region. Copenhagen, Denmark: WHO.
Peterson, B. J., & Fry, B. (1987). Stable isotopes in ecosystem studies. Annual Review of Yiannas, F. (2018). A new era of food transparency powered by blockchain. Innovations:
Ecology and Systematics, 18, 293–320. Technology, Governance, Globalization, 12, 46–56.
Post, D. M. (2002). Using stable isotopes to estimate trophic position: Models, methods, Zhang, X., Liu, Y., Li, Y., & Zhao, X. (2017). Identification of the geographical origins of
and assumptions. Ecology, 83, 703–718. sea cucumber (Apostichopus japonicus) in northern China by using stable isotope ratios
PricewaterhouseCoopers (2016). Fighting $40bn Food Fraud to Protect Food Supply. Vol. and fatty acid profiles. Food Chemistry, 218, 269–276.
2017. PricewaterhouseCoopers. Zuo, R. (2013). ITRAX: A potential tool to explore the physical and chemical properties of
Primrose, S., Woolfe, M., & Rollinson, S. (2010). Food forensics: Methods for determining mineralized rocks in mineral resource exploration. Journal of Geochemical Exploration,
the authenticity of foodstuffs. Trends in Food Science & Technology, 21, 582–590. 132, 149–155.
Rehbein, H. (2008). New fish on the German market: Consumer protection against fraud
304