Article is available in full to IFST members and subscribers.

Register on the FST Journal website for free

Click the button to register to FST Journal online for free and gain access to the latest news


If you are an IFST member, please login through the Members Area of the IFST website. 













Trace the food, not the packaging!

Laurie M. Clotilde, Anthony Zografos, and Molly A. Trump of US-based SafeTraces describe the performance of a new technology for tracing food in the supply chain using DNA barcodes that can be applied directly to the food.


Globalisation of food sourcing and political and commercial realities require increasing supply chain transparency, accountability and security. One key to achieving gains in all three areas lies in the ability to trace the source of foods and their ingredients, from fork to farm.

Today, case level traceability entails a complex system of handoffs along the supply chain from producer to packer, distributor, retailer and, ultimately, the consumer.

Printed barcodes on the packaging identify the product lot. But once food, and particularly fresh produce, reaches retail, it is usually removed from the original packaging and traceability is generally lost. Case level traceability traces the packaging and not the food. In the event of a post retail quality or safety problem, such as a foodborne illness outbreak, traceback investigations are frequently inconclusive or take several weeks to complete[1].

Traceability improvements would benefit: (a) the regulators by reducing the resources required to complete investigations; (b) the public, since reduction of the duration of traceback investigations is key to the containment of quality or safety problems, and (c) producers by facilitating the ability to isolate the source and extent of safety or quality issues, thus improving the implementation of corrective and preventive actions and minimising the delay and scope of the recall and associated liabilities.

US-based SafeTraces has developed a patent-protected, natural, edible, odourless, tasteless, on-food traceability solution. Advances in bio-engineering have produced a material that enables the development of an efficient, effective, and low cost system to trace the food, not the packaging.

This material is a combination of short DNA sequences (i.e., <100 base pairs), which can be either synthetic or genomic DNA drawn from organisms (e.g. Thermotoga maritima) that are not expected to be present in the food environment[2].

These DNA sequences have already been recognised by the United Stated Food and Drug Administration as Generally Recognized As Safe (GRAS). These sequences, referred to as tracers or tags, can be used to form unique combinations, called DNA barcodes.

This DNA barcoding method is different from the taxonomic method that uses a short genetic marker in an organism's DNA to identify it as belonging to a particular species.

The Safetraces approach employs 64 distinct sequences, with each sequence representing a specific bit in a 64-bit set; presence of the sequence sets the bit to 1, and absence of the sequence sets it to 0. By employing only 64 sequences, 264 unique combinations are created, referred to as ‘DNA barcodes’. An advantage of this approach is that it can analyse 264 unique barcodes by employing only 64 probes and sets of primers for detection via the Polymerase Chain Reaction (PCR) process. This makes it highly scalable and sets it apart from other DNA tagging methods which, in practicality, can only be used for authentication or identification but not traceability, because of the limited number of unique sequences they could economically support. A second advantage is that these barcodes can be applied directly in or on the food, thus allowing tracing of the product even if it has been removed from its original packing.

In its simplest implementation, a unique combination of DNA sequences may be used to identify a grower or producer. In a standard implementation, a unique DNA barcode can identify a single product lot. In an even more complex implementation, multiple 64-tag sets may be used to form multiple barcodes, each associated with a specific step in the supply chain. A single piece of produce might bear many different barcodes. Figure 1 shows how the first DNA barcodes may be used to identify the grower, the field, the picking crew, the machine used and the date. The second barcode can identify the packer, the packing line, the packing date and the ship date. These two barcodes may coexist on the same piece of produce and can be read independently.

Application of the DNA barcodes is very simple. They may be added directly to various coatings already in wide use, such as carnauba or other waxes (e.g. pomme fruits, citrus, stone fruit), silicone oils (e.g. tomatoes, tropical fruit), sprout inhibitors (e.g. potatoes), lipid-, polysaccharide- and protein-based edible coatings. DNA barcodes encapsulated in maltodextrin, salt, starch or other material may be used for dry goods, such as beans and cereals. DNA barcodes may be also added directly to liquid goods, such as juices, milk and oils. The DNA barcodes are added to food in parts per billion or less. If they were applied on every food a person eats, over the course of a lifetime, we estimate that it would add up to just over one gram of material.

The most common concerns regarding DNA barcodes are their stability over time and their integrity when items with different DNA barcodes are mixed (commingled), which presents potential for ‘cross-contamination.’

Figure 1: Examples of two 64-bit DNA barcodes (blue and green) implemented on two supply chain

In its simplest implementation, a unique combination of DNA sequences may be used to identify a grower or producer.’

DNA barcode stability and integrity

In a laboratory experiment, different varieties of fruits were coated with carnauba waxes containing a combination of up to three different tracers, effectively forming DNA barcodes. The amount of wax applied was equal to the rate of commercial applications (i.e. 1 L of wax per 1000 kg of product).

The fruits belonged to different varieties of apples and citrus. They had the following DNA barcodes represented as 0 for absence or 1 for presence of Tracers 1, 2, and 3: 000 for the Red Delicious apples, 100 for Green apples, 010 for Yellow apples, 001 for Jazz apples, 110 for Fuji apples, 011 for oranges, and 111 for lemons. The fruits were commingled in a basket, stored at room temperature, and tested over time on days 0, 4, 5, 7, 11, 20, 21, and 26.

Each fruit was washed individually and swabbed using a dry cottontipped swab. The swab was suspended in a buffer and tested on the Chai Open qPCR system using an internally developed 16-min protocol. The total test time was <18 minutes.

The results (Figure 2) are presented as Cycle Threshold (i.e., Ct; unit of measure of qPCR technology) with the lower values representing higher levels of the tracers[3].

A threshold of Ct=29 has been established to differentiate positive from negative signals, meaning that a Ct<29 sets the corresponding bit to 1 and a Ct>29 sets the corresponding bit to 0. The fruits that had been originally tagged with the DNA barcodes remained positive for the entire duration of the study (i.e., 26 days). This is generally considered longer than the shelf life of both apples and citrus when stored at room temperature. There is some inconsistency in the measurements on a day-to-day basis, which is likely to be due to normal assay variations, including the swabbing of the produce (i.e. difference in force, duration, etc.). Each fruit was washed before each test and the results suggest that washing does not significantly accelerate the degradation of the barcodes. In fact, the level of tracers on the fruits remained fairly constant for the duration of the experiment.

A trend is more apparent with the ‘negative fruits’ (i.e., those that had one or more bits set to 0). Over time, there is increasing transfer of the DNA barcodes from the positive to the negative fruits, even though both negative and positive fruits were washed prior to each measurement. However, all negative fruits remained negative for the duration of the experiment.

In both cases (i.e., negative or positive), there does not appear to be any dependency on the type of product (i.e., apple or citrus) or the variety.

The experiment identified with 100% accuracy the DNA barcodes on commingled fruits, which were stored at room temperature in excess of their expected shelf life.

Figure 2: Stability of DNA Barcodes over time under commingled conditions

The experiment identified with 100% accuracy the DNA barcodes on commingled fruits, which were stored at room temperature in excess of their expected shelf life.’

Commercial implementation on an apple packing line

The DNA barcodes were also applied on a commercial apple packing line (Figure 3).

An off-the shelf auto sampler (Figure 3, top) creates unique DNA barcodes that can be used to identify product lots. The auto sampler, or dispenser, is loaded with vials containing each sequence and is used to create unique combinations (i.e., mixtures) of these sequences (i.e., DNA barcodes). The dispenser is connected to a main database, which ensures that each barcode is unique. The mixtures are then injected in the carnauba wax stream at a rate proportional to the wax flow rate. For the injection, a simple dosing system is used (Figure 3, middle).

The results from a pilot implementation are presented in Figure 4. In this case, the DNA barcode was introduced in the entire wax tank. The purpose of the test was to evaluate the transition time from the introduction of the DNA barcodes in the wax tank to their actual detections in both the wax and on the fruit.

Under the standard commercial implementation, and because of the proximity of the injection system to the nozzle system, the transition time was expected to be in the order of seconds as opposed to minutes.  One minute after the DNA barcode was introduced in the wax tank, both wax and apples appeared positive for the barcodes.

Over the course of the ensuing 24 hours, the concentration in the wax decreased slightly. This decrease may be attributed to assay variations or degradation of the DNA in the liquid wax environment. Laboratory tests have shown that some degradation does occur through long-term exposure (i.e. several weeks) of the DNA to the liquid wax.

However, in a commercial implementation, the DNA barcodes are maintained in food-grade ethanol and are delivered by the injection system to the wax line just before the spray nozzle. The laboratory tests cited above have demonstrated that the DNA barcodes are stable in ethanol practically indefinitely (i.e. no degradation observed after several months of exposure). As a result, in a commercial implementation the exposure of the DNA barcodes to liquid carnauba wax is reduced to a few seconds.

Similarly, a small decrease in the presence of tracers on the waxed apples was observed over the course of the ensuing 24 hours. This decrease may be due to the decrease in the DNA concentration in the wax or normal assay variation.

However, the apples remained below the positive threshold Ct=29. It should be also noted that a full tank of liquid wax is normally used over approximately 3 days and during such a relatively short time period, the DNA concentration in the wax remains well within the range required to produce apples below the Ct threshold.

Figure 3: Commercial implementation on an apple packing line
Figure 4: Simple DNA barcoding on a commercial packing line.

The technology is very simple to integrate into the apple packing process’


The feasibility of using DNA barcodes for traceability in the food supply chain has been demonstrated. The DNA barcodes are stable for the shelf life of the produce. Cross-contamination does not seem to be an issue when produce with different barcodes is commingled as it is not sufficient to cause errors in the identification of the DNA barcodes.

The technology is very simple to integrate into the apple packing process (which is very similar to those of most pomme fruit, stone fruit and citrus) and ease of detection has been reduced to <18 minutes.

Laurie M. Clotilde, Anthony Zografos, and Molly A. Trump, SafeTraces, Inc., Pleasanton, California, USA

Web: Email:


  1. Hoffmann S, Maculloch B, Batz M. 2015. Economic Burden of Major Foodborne Illnesses Acquired in the United States. United States Department of Agriculture. Economic Information Bulletin 140.
  2. Jackson S, Rounsley S, and Purugganan M. 2006. Comparative sequencing of plant genomes: choices to make. Plant Cell. 18(5): p. 1100-4.
  3. Pohl, G; Shih, leM (2004). "Principle and applications of digital PCR". Expert Review of Molecular Diagnostics 4 (1): 41–7.

View the latest digital issue of FS&T or browse the archive


Click here

Become a member of the Institute of Food Science and Technology


IFST Twitter Feed