To project overview

Analysis of Complex Biological Mixtures by a Novel Method: “DNA-SHAPES” - A Model for Artificial Taste

Semper Ardens Research Project| 03/08/2016

Food and beverages are extremely complex mixtures of substances that upon interaction with thousands of taste receptors on our tongue send a cascade of signals to our brain to create our perception of taste. We will in the DNA-SHAPES project endeavour to mimic and go some way towards understanding this mechanism by mixing the food with billions of nano-sized biosensors, which simultaneously can probe essentially all substances in the food and transform these associations into a digital output in the form of a large set of DNA sequences. Using new high speed DNA sequencing methods and computer analysis, we hope to be able to develop ‘artificial tasting machines’ that not only can help determine the sensory characteristics but also enable assessment of quality, authenticity, and contamination in food in a rapid manner for use in industrial production. We imagine that this DNA sensor-based characterisation of food constituents will help us to understand the complexity and specificity of human taste perception and how to mimic it. The DNA-SHAPES method also has additional potential in grand solution applications, for instance, for determination of the effects of global environmental pollution and for the diagnosis of human diseases based on analysis of various body fluids.

Human sense of taste: Humans have about 5000 taste buds, each consisting of about 100 cells. Each cell expresses 50-100 different receptors, which upon binding of specific tastants initiate intracellular signalling cascades transducing into neuronal signalling to the brain. In essence, this means that the complex signal from approximately 500.000 cells is creating the taste conception in our brain.

‘Artificial Tasting Machines’

Animals and humans are equipped with sensory organs for taste mainly located on our tongue. The taste impression is created when substances bind to taste receptors on cells located on our taste buds. Upon binding of food ingredients to the receptors an intracellular signalling cascade is initiated transducing into neuronal signalling to the brain (1).

The aim of this proposal is to mimic the natural concept artificially by creating a library of billions of different molecular biosensors that specifically can bind to individual substances in solution. By selecting the subclass of modules that can bind to the components in a particular food mixture and subsequently characterise them Kjems and his colleagues hope to create a “picture” of the ingredients - very similar to the principle applied by our brain. However, their sensing strategy will not use protein receptors but rather more than 1016 free-floating DNA molecules folded into a complex set of 3D structures (2). Professor Jørgen Kjems commented:

“At first I thought that this was a farfetched idea but when I started calculating the numbers I realised that it should be feasible with new high speed DNA sequencing techniques”.

How can DNA sense food ingredients? Double stranded DNA usually constitutes the carrier of our genetic information, but if made single stranded it can in fact form billions of 3D shapes that can bind specifically to other molecules. Such DNA modules are also called aptamers and will be used as the sensory mechanism in our method (3). The advantage of aptamers is that they are very small sensors. In fact, more than 1016 different sensor DNA molecules can be made in one reaction and be contained in one teaspoon of water. Upon mixing with the food the DNA sensors can interact with a wide range of targets including proteins, peptides, sugars, and other biological molecules (4). The novel idea by Kjems and his colleagues is to use billions of aptamers in parallel – something that only has become feasible after the recent invention of next generation DNA sequencing techniques where a million of DNA building blocks can be decoded in a few seconds.

Aptamers (from the Latin aptus - fit, and Greek meros - part) are short strands of DNA (or RNA) that can adopt 3D shapes, which in a specific manner can recognise a wide range of targets including proteins, peptides, sugars, and other biological molecules.

Proof-of-Concept, Why Not in Beer?

The method will first be tested on beer, which is from a human sense perspective infinitely complex and differentiated in terms of ingredients and tastants linked to the natural processes the ingredients undergo during fermentation. Vice President Birgitte Skadhauge who is dealing with beer components in the brewing process at the Carlsberg Research Laboratory explains:

“If the method is working as predicted it will give us more detailed insight into the content of beer. The interesting part will be the degree to which the fingerprints of the ingredients can be calibrated to human perception-based interpretation of foods and beverages, as well as allow a prediction of freshness, quality, authenticity, adulteration, and the presence of allergens or contaminating microorganisms and other unwanted ingredients.”

Correlating the DNA SHAPES Fingerprints with the Human Senses

The DNA-SHAPES team:

The participants in the DNA-SHAPES include:
•Jørgen Kjems (main applicant), professor and director, iNANO and Department of Molecular Biology and Genetics
•Derek Victor Byrne, Professor and Science Team Leader of Food Quality Perception & Society (FQS) at The Department of Food Science (FOOD), Aarhus University
•Jan Enghild, professor at iNANO and Department of Molecular Biology and Genetics.

To address how the results can be correlated with human perception of taste Professor Derek Victor Byrne and his science team (Food Quality Perception & Society) from AU-FOOD will provide the human senses expertise and lead the perception measurements in the project (5,6). Quantitative sensory profiling will be performed on fresh and age accelerated beer samples in parallel to DNA-SHAPES analysis. The effects of different malt, yeast hops, and adjunct types will also be characterised with respect to brewing process variables. DNA-SHAPES fingerprints will be linked to these sensory profiles and aroma analyses to build benchmarking and calibration models for future quality measurement instrumentation. Such ‘artificial taste sensors’, calibrated and benchmarked by direct predictability and causality to human sensory measurements as a direct measure of quality, have not been realised before. Extrapolating from this, such taste sensors could also potentially be revolutionary and ground-breaking in the field of human sensory perception in relation to shedding light on the specific mechanisms underlying how we taste. This has far reaching consequences for human food perception, eating and health. Derek Byrne says:

“When Jørgen and I discussed the idea of DNA based specificity for food and beverage quality defining constituents, I was intrigued by it as a solution for quality measurement, but thought let’s take this one step further and try to see if this can also help us understand how we taste. Axel and Buck received a Nobel Prize in 2004 for elucidating how our smell system works, but taste is still very much undiscovered country”.

DNA-SHAPES can definitely shed light on taste, a most basic but complex human sense.

Potential Applications

The primary goal of this project is to create a model for human perception of taste and a rapid analysis tool for food quality designation. Both these accomplishments will have an enormous potential for the food and beverages industry in terms of allowing rapid and accurate assessment, which is not possible at present.

With new upcoming miniature devices for DNA sequencing it is foreseeable that we within a few years can create hand-held devices for rapid and accurate assessment of food-quality parameters, -authenticity, -contamination, and -designation. This is of high relevance for e.g. general screening of the molecular content of imported and exported goods, real time quality assessment in production lines, and tracing of contaminants.

However, the DNA-SHAPES technique can readily be expanded for a wider range of applications. It will e.g. have obvious applications in pollution control of water and soil as well as in healthcare, where characterising the mixture of proteins and other biomolecules in biofluids (e.g. blood, urine, saliva, faeces) can help disease diagnosis.

The “Tasting” principle in the DNA-SHAPES method is here exemplified with beer. A large number of nano-meter sized biosensors are synthesised in a test tube and added to various types of beer. Biosensors that are triggered by beer ingredients are separated from non-responsive variants and characterised in the form of a large set of DNA sequences. By aligning the data with human taste perceptions we can “teach” the computer to recognise specific ingredients and taste.


1. Barlow LA: Progress and renewal in gustation: new insights into taste bud development. Development. 142:3620-9 (2015)

2. Andersen, E.S., Dong, M., Nielsen, M.M., Jahn, K., Subramani, R., Mamdouh, W., Golas, M.M., Sander, B., Stark, H., Oliveira, C.L.P., Pedersen, J.S., Birkedal, V., Besenbacher, F., Gothelf, K.V., and Kjems,J.: Self-assembly of a nano-scale DNA box with a controllable lid. Nature 459:73-76. (2009)

3. Sun et al. A Highlight of Recent Advances in Aptamer Technology and Its Application. Molecules. 20:11959 (2015)

4. Dupont, D.M., Larsen, N., Jensen, J.K., Andreasen, P.A., Kjems, J.: Characterisation of aptamer-target interactions by branched selection and high-throughput sequencing of SELEX pools. Nucleic Acids Res. 43:e139 (2015).

5. Dijkterhuis, G., and Byrne, D. V. Does the mind reflect the mouth? The future of sensory profiling. Critical Reviews in Food Science & Nutrition. 45:527-534 (2005).

6. Mielby, L. H., Andersen, B. V., Jensen, S., Kildegaard, H., Kuznetsova, A., Eggers, N., Brockoff, P., Byrne, Derek V. (2016). Changes in sensory characteristics and their relation with consumers' liking, wanting and sensory satisfaction: Using dietary fibre and lime flavour in Stevia rebaudiana sweetened fruit beverages. Food Research International, Vol. 82, 04.2016, s. 14-21.