‘They Live’

Our Food Lab is housed in our Center’s former canteen kitchen, which we took over and began refitting in summer 2022. After we had mostly finished setting it up, in summer 2023, the Center decided to refurbish the eating area next to the kitchen.

Part of this project involved swapping the former metal rollers for clear ones to let more light into the eating area. When this happened, we needed a way to partially shield the lower rollers to cover the backs of the equipment and to find the right balance of privacy and openness between the Food Lab and the eating area.

Instead of making something merely functional, we used this as an opportunity to create an artwork that can showcase the food innovation work being done in the Food Lab, and enrich the environment of our Center.

We selected five key microbial species we work with in the Food Lab, one for each slat, and turned the universal identification gene (16s rRNA or ITS) for each species into binary code, yielding a unique ‘barcode’ pattern.

The species are, from top to bottom:

Microbe	Primary use
Aspergillus oryzae	The main kōji fungus, which we use in many of our experiments.
Bacillus amyloliquefaciens	A common bacterial symbiont of A. oryzae, often found in kōji, miso, and associated plant-based umami products.
Penicillium camemberti	An important fungus for making bloomy-rinded cheeses, which we use in our plant cheese work.
Lactiplantibacillus plantarum	A key lactic acid bacterium involved in the lactic fermentation of plant substrates.
Rhizopus oryzae	One of the main fungi we use to create meaty flavours from side-streams.

These 16s rRNA and ITS sections of DNA exhibit minimal inter-species variations, making them useful identifiers or barcodes’ for species identification in the lab. Visualising these DNA sections is almost like unveiling the species' unique fingerprint, and representing this genetic signature as a barcode naturally aligns with its role as a biological identifier. Here we share a bit more about how we translated these DNA sequences into the barcode patterns we used.

DNA encodes information using four nucleotide bases: A, T, C, and G. This quaternary system can be translated into a binary format, which could then be represented visually through opaque (black) and transparent (white) bars. However, this conversion increases the number of elements needed to encode the same information. Specifically, in this barcoded version, each nucleotide requires two bars (Table 1).

Table 1: Translation of nucleotides to barcode representation.

To streamline the encoding process and reduce its length (to fit the gene’s entire DNA sequence on the window), we compressed the encoding back into one element per nucleotide. One approach would be to introduce two new bar designs, such as patterns like stripes and polka dots, making the encoding a simple one-to-one matching—one bar pattern, one nucleotide. Another method was to merge the initial two-bar nucleotide representation into a single bar (Table 2). This approach allows for a more scalable data representation compared to the nucleotide-to-pattern matching approach. By providing an efficient 2D structure for the translation of the DNA, it makes it easier to further reduce the length of the translation by stacking more elements into one column. The barcode becomes 2-dimensional as the information flows in two directions: up to down, and left to right. The structural elements of the design, formerly bars, now become pixels.

Table 2: Translation of nucleotides to column representation.

This 2D structure can be visualised as a stacked representation of the nucleotide sequence, where each column encodes a segment of the DNA sequence (see Table 3 for a comparison of the three approaches). This method maintains the fundamental characteristic of DNA—a linear sequence of nucleotides—while providing a compact and visually distinct representation.

Table 3: DNA sequence represented with different methods.

With the translation from DNA to visual design completed, we iterated over several parameters to refine the result. The initial parameter was the straightforward mapping of nucleotides to their respective patterns, aiming to strike a balance between opacity for privacy and translucency for light. Different mappings gave slightly different aesthetic impressions (Table 4). 

Next, we experimented with the size of the pixels. We explored two contrasting approaches. The first approach scaled the pixel size from the marker gene’s sequence so that the barcode would cover the entire window. The second approach defined a fixed pixel size calculated to fit best into the windows’ slightly different dimensions and extends the DNA sequence beyond the gene of interest, as far into the genome as necessary to fill the same amount of space. We opted for this second approach because then the pixels would be consistent across the five slats, giving a more coherent overall impression.

Table 4: Different results obtained for the same DNA sequence based on different mappings of nucleotide to binary pixel code.

At this stage, we installed this first prototype in the canteen. We then quickly realised the transparent pixels allowed too much visibility into the kitchen. This compromised the project's main objective of providing greater privacy.

To address this issue, we introduced a new parameter: overall opacity, defined as the percentage of opaque pixels in a sequence. After evaluating all possible configurations, we concluded that we needed to adopt a new barcode paradigm based on a 4-pixel mapping (see Table 5). Though this doubled the length of the barcode, this wasn’t a concern as we had a lot of window to cover. Experimentation revealed that incorporating spatial symmetries into the pixel mapping yielded distinctive barcode patterns. Specifically, mapping 5A, which employs horizontal symmetry, produces horizontal patterns with diagonal accents, whereas mapping 5C, based on diagonal symmetries, generates a checkerboard-like pattern. These are quintessential examples of emergent properties.

We tested three final prototypes, comparing their performance. Ultimately, we chose the mapping 5B with the highest opacity (75%), as it not only maximised privacy but also effectively conveyed the barcode's aesthetic when applied to the transparent surface.

Table 5: The 4-pixel paradigm, allowing more control over the patterns within the barcode.

The final installation is shown below in Figures 1 and 2.

Figure 1: Our research assistant Mathieu, who created the binary code design for the project, standing in front of the final installation.

Figure 2: A close-up of the five panels. We’re pleased with how the installation allows natural light to shine through, whilst affording privacy to the people working inside.

As for the title. They Live is a 1988 American science fiction action horror film written and directed by John Carpenter, based on the 1963 short story ‘Eight O'Clock in the Morning’ by Ray Nelson. The film follows a drifter who discovers through special sunglasses that the ruling class are aliens concealing their appearance and manipulating people to consume, breed, and conform via subliminal messages in mass media. Since its release the film has acquired cult status and entered pop culture, notably having a great influence on the street art of Shepard Fairey, who took from the film the name of his streetwear label, Obey.

There are many comparisons to be drawn between the film's themes and our growing knowledge of the omnipresent microbes we share space with. The holobiont concept, in which humans and other creatures are recast as symbioses of ‘host’ and microbiota, raises not only new scientific questions, but also new, and indeed age-old, social and ethical ones: for example, how much authority do we have in making daily decisions, and how much is influenced and driven unseen by these tiny co-habitants of our bodies and the surrounding world? And who even is this ‘we’ when we pose such a question?

As we grapple with the answers and their implications across fields, one thing is for sure—whoever these beings are, they live.

Contributions & acknowledgements

Josh proposed to make an artwork. Kim and Josh conceptualised the idea. The members of Sustainable Food Innovation selected the five species to use. Mathieu did the coding work to render the 16s and ITS genes as barcode images. Mathieu, Kim, and Josh iterated the designs. Josh and Kim worked with Gabriel Szima, Facility Manager at our Center, to prepare the work, which was printed and installed by Dansign. Mathieu wrote the first draft of this article, with editorial contributions from Josh and Eliot.

Related posts