Using this technique, Tabor and his colleagues at UCSF and the University of Texas, Austen created ghostly pictures of squid and people in 2005. The images were a very high resolution, with each bacteria representing one pixel.
Tracing the edge of an image is more complicated.
Instead of responding to a physical signal -- the lack of light -- the bacteria also have to respond to a chemical cue from surrounding bacteria.
Using viral vectors, the scientists injected genes into the bacteria that, when triggered by a beam of light, causes light sensitive bacteria to emit a chemical that turns the surrounding, non-lightened bacteria black.
The interaction of the two signals creates a pencil-thin line through a plate of agar.
This means only cells right on the boundary of light and dark, can create the dark pigment to trace the image.
"If you are in the dark you can make the signal but can't listen to it," said Tabor, "And if you are in the light you can't make the signal but you can hear it."
The research is a good example of how a bacterial computer could operate. Each E. coli cell detects light or detects the chemical cue at the same time and processes the information simultaneously to create an edge. If the bacteria were a computer, this would be called parallel processing.