Researchers design organisms using computer simulations, frog cells

Scientists from Tufts University and the University of Vermont (UVM) successfully created programmable, living nanobots from the cells of frog embryos, called xenobots. Michael Levin, the Vannevar Bush Professor of Biology and director of the Allen Discovery Center, and Dr. Douglas Blackiston, a scientist in Levin's lab, co-authored the report published on Jan. 13 with their UVM counterparts.

The research project was composed of two parts.

Levin and Blackiston conducted the biological portion of the research. Dr. Josh Bongard, director of the Morphology, Evolution & Cognition Laboratory at UVM, and Sam Kriegman, an evolutionary robotics Ph.D. candidate at UVM, led the computer science part of the project.

The project focused on finding a way to design organisms from frog cells based on a computer algorithm and artificial intelligence, according to Blackiston. He added that the computer program used by Bongard and Kriegman simulated task-specific cell configurations, selecting the best performer of the group.

Blackiston explained that the repeated simulations, which were used to replicate the configuration with organic material from the frog cells, were similar to evolution.

"Over many different generations, just like real evolution, you come up with something much better at solving the task,” Blackiston said.

Levin explained that the study's conclusion was that cells with normal frog genomes are able be reassembled in a completely different configuration with different behaviors, without any genetic manipulation.

“What the skin cells did was assemble a network that [made cells] able to talk to each other and pass the electrical signals that get the cardiac cells contracting, and when the cardiac cells contract it generates force,” Levin said. “The whole point of the computer simulation was that we were able to see under those circumstances how to sculpt the thing a little bit.”

Levin said that he worked to define the research direction, conduct and design the project while Blackiston developed the protocol. At UVM,Bongard made the simulator to evolve robot designs and Kriegman designed the code for the virtual evolution.

Vice Provost for Research Caroline Genco, whose office evaluates funding proposals based on their impact, cross campus interactivity and feasibility over a three-year period, said that Levin’s research was selected to receive funding because it was innovative and it had great potential for future study.

“[Dr. Levin’s] work is really one of the pillars of biological research in [the School of Arts & Sciences]. He is one of our stars,” Genco said. “He’s got a fantastic group — they’re probably some of the most innovative scientists we have."

Blackiston explained that scientists can use the xenobots to explore questions about computer-designed organisms made from organic material.

“We’re trying to create this sandbox of development where different groups can see what they can use the xenobots for and build the organisms specific to their purposes,” Blackiston said. “This is really a great testbed for sort of anybody that is interested in this field to ask these sorts of questions using the existing technologies.”

Levin noted that he is working on analyzing other data focusing on modifications of the xenobots, as well as their different behaviors and abilities, and expects another paper published this year.

Levin emphasized that this research helps begin to answer much broader and more fundamental questions about cell biology.

“I think it’s important to put this in the context that the big thing here is not just having robots made of frog cells, but that this is really a computational question about how cells know what to do and how they make decisions,” Levin said.

Levin elaborated, connecting his ongoing research to other unanswered research questions in computer science.

“It’s still pretty unknown how [cells] know how to make a particular structure, so this is part of that really large question that’s at the center of all of biology and also of a lot of computer science and robotics, on how collectives make decisions,” Levin said.