FEB 25, 2020
The authors describe an evolutionary algorithm that designs biological machines—an approach that could one day possibly be used to safely deliver drugs or repair damaged tissue inside the human body.
The group developed a computational algorithm to design biological machines to perform specific tasks such as pushing an object across a petri dish. These “machines” were modeled with 2 basic building blocks: a structural cell and a contracting cell. Each design then underwent simulated examination, evaluated on its ability to perform the set task. The best design predictions were then physically built by combining different biological tissues using cells from Xenopus blastula (e.g., a structural cell and contracting cell).
Using microsurgical tools, scientists could dissect Xenopus blastula into the shape predicted by the algorithm, overlaying certain areas with contracting Xenopus cardiomyocytes. These organisms were then observed for spontaneous behaviors. They used these designs to validate simulated predictions.
Xenopus blastula are a hardy tissue with significant plasticity. Whether human derived pluripotent stem cells will be compatible and capable of similar behaviors is not a simple question.
These "designer" organic robots may be one day employed in patients to perform tasks such cleaning up the trabecular meshwork or repairing demyelinated neurons.
Image credit: Sam Kriegman, The University of Vermont