This year the Royal Swedish Academy of Sciences has decided to award Osamu Shimumora, Martin Chalfie and Roger Y. Tsien for the discovery and development of the green fluorescent protein (GFP). This is what is being used in the bio.display project as well. Many thanks to the above people to make my work possible!
Allison Kudla is working with similar concepts, using biological organisms in her artwork. Her work seems to concentrate on using living organisms as a raw material for creating a work, but is still very interesting, especially this one.
Eva Jablonka published an article titled “Epigenetic learning in non-neural organisms“, basically suggesting that single-celled organisms, such as bacteria, can be trained in a way we associate to organisms with a nervous system. This would imply that to some extent the bacteria can be made to “learn”, and then use that knowledge later on in similar situations.
The article is available online in PDF form here, of course there are references abound.
Here’s a visual illustration of the process itself:
getting a 16×16 bitmap into our dropdrawer program, in this case the ChickenRunner’s icon from the age-old computer game Berzerk
letting the fab@home fabber drop arabinose on the plate
wait overnight, and see the results
The experiments done from yesterday to today show a clear result that the size of arabinose drops really matter, and that the correspondence between the brightness of the result and the size of the drop is not linear. It seems that a drop size of at least 0.02 is preferable (say a pushout of 0.03 and a suckback of 0.01), sometimes 0.015 works as well. It’s also visible that we can’t get better than a distance of 5mm between the drops, but 6mm seem to work out best. Shorter distances get along better with smaller drop sizes, probably because the liquid flows together. Bigger drop sizes also help in the process as they stick to the agar gel easier.
The following prints resulted in visible output (spacing / drop size / suckback):
Whereas the following prints resulted in practically invisible results (spacing / drop size / suckback):
The following image shows three prints of spacing 6, with drops sizes below (spacing / drop size / suckback):
The following images shows prints of spacing 5, with drops sizes below (spacing / drop size / suckback):
The following image shows three different prints of different sizes (7, 6 and 5) and similar drop sizes, see below (spacing / drop size / suckback):
I’ve made a number of prints today, basically experimenting with different spacings between the drops. All images were 16×16 pixels, with the following parameters:
we’ll see tomorrow how they turn out. Unfortunately leveling is really an issue, especially with small drops: the syringe touches the surface of the agar on some parts of the plate, but not on the others. So drops are not placed everywhere where they should, and in some places, accumulated drops are made.
Yesterday I made a number of images to try out different resolutions and drop distances, to see which work the best. One limitation is the size of the droplet that can be reliably dropped by the fab@home 3D printer. Another is that the arabinose dropped will dissipate through the agar, in effect creating a much larger pixel than the drop itself. Also, the bacteria can overdose on the arabinose, in the end not displaying anything.
I experimented with a simple X pattern, and also with some space invaders:
The initial resolution was 16×16 pixels, with various drop sizes. The drop sizes are defined by the amount of downwards movement that is made by the syringe tool, pushing the liquid out. The spacing between the drops can also be played with, and thus I did.
The three invaders below all have 5mm between the drops, and the drop sizes are also the same, 0.03. The difference is only in the amount of suckback after a drop is made - that is, some liquid is pushed out, and then there’s a backwards movement in the syringe, sucking excess material back. The invader on the left on both images has a suckback of 0.01, while the ones on the right have none and 0.015, respectively.
It seems that no suckback resulted in an overdose of arabinose (see that the silhouette of the invader is visible, but the inside is dark). It also seems that a suckback of 0.015 was too much, though this is unclear why.
I also experimented with 32×32 pixel resolution images, with shorter distances between the drops, and smaller drops still. Here, the drops really got together during printing (see yesterday’s post), thus the result is really just blog blobs, with arabinose overdose in the middle:
The invaders, from left to right, were made with the following settings:
The next experiments will involve more spacing and smaller drops, so as th prevent arabinose overdose, and also to enhance the effect of single pixels.
It’s also a challenge to make these photos, in low light conditions it’s very difficult to focus properly. Still, thanks to Dániel Molnár for his tripod, as my own is still missing 
Yesterday’s initial tests with the printer seems to have worked out fine. The rudimentary figures I tried to draw show up from the plates. Open questions about preparations for today were also answered by Maarten.
One of yesterday’s open questions was the amount of ampicillin that should be added both the liquid culture and the LB agar. It’s clear now - the final concentration should be 50 μg/ml, and thus I have to calculate the 250mg/ml concentration of ampicillin accordingly. Thus, the preparations I’ve done today included, for the liquid culture:
For the plates I made:
Let’s hope that the way too low amount of ampicillin added yesterday still makes things work for now, as I’ve printed all 10 plates today, that were prepared yesterday. Another risk is that I might have mixed the bacteria with the LB agar while it was too hot. Unfortunately there’s no other way than to guess by hand if the LB agar is cold enough…
Today I tested printing with a wide variety of drop sizes and drop distances. We’ll see tomorrow which ones work out the best, if the plates are any good 
I’m back in Leiden, with the aim of putting together the modified fab@home 3D printer and the plates of bacteria in the lab. The following 2 weeks will be the moment of truth, and thus we’ll see if he project was worthwhile.
A lot of preparation has already been kindly done by Maarten de Smit, who has arranged for the production of large quantities of LB agar, with 0.75% agar content (this is the agar content ratio of what is called the ‘top agar’). Trudie Bowers has provided me with all sorts of equipment and materials at the lab, like arabinose, liquid LB, pipettes, Erlenmayer flasks, liquid water bath, gas torch, large pipettes with rubber pumps, etc. Anne Kienhuis is organizing everything for me very kindly.
I will be working with 12×12cm plates to create my images.
The daily routine here will consist of maintaining a stable supply of plates full of bacteria in LB agar, so I can put them into the printer, and then experiment with the plates in the printer. To provide a stable supply, Maarten has grown some bacteria, taken from the cryogenic storage and spread out on a plate. Using this bacteria, overnight liquid cultures have to be made each day. Using these liquid cultures, the plates can be made themselves with bacteria, which also take overnight, and have to be put into an incubator.
Today I made a liquid culture with 100ml of liquid LB and some ampicillin. Here I’m in a bit of a trouble, as the ampicillin bottle says 250mg/ml, which seems an extremely high concentration, about 5000 times more concentrated than the 50μg/ml carbenicillin I used to use. Thus I put 1μl of it into 100ml of liquid LB, which should be 250μg of ampicillin. And of course a scrape of bacteria from the plate with the spread out bacteria.
I also made 10 plates, 12×12cm each with LB agar and liquid cultures. Used 1 liter of LB agar (0.75% agar), and put 50ml of liquid culture into it. I had the same dilemma here with the ampicillin, so these guys also got only 1μl, which should be 250μg again. I wonder if it works.
Afterwards, I started to fire up the printer and print on plates prepared by Maarten. At first of course there were some problems, but at the end it seemed to work out. Well, we don’t know of course - I got reasonably small drops of arabinose, but we’ll have to see tomorrow if they work out. Now I’m using a 16×16 pixel grid, which is old-school icon size.
Heather Ackroyd & Dan Harvey have been creating images by growing grass selectively since about 2001. Recently they’ve created a high-impact piece related to the Wimbledon Tennis tournament.
A nice form of bio-art.