Scientists have developed a robotic artificial intelligence system to self-determine the optimal conditions for growing the replacement retinal layers needed for some treatments aimed at restoring vision. In the latest experiment, the system underwent a trial-and-error process covering 200 million possible configurations and managed to dramatically improve the viability of cell cultures needed for regenerative medicine therapy. This achievement is a good example of how the automated design and execution of scientific experiments can increase the efficiency and speed of research in fields such as biology.
Research in regenerative medicine often requires many experiments that are time-consuming and laborious. Specifically, creating specialized tissue from stem cells (a process called induced cell differentiation) takes months of work, and the degree of success depends on a wide range of variables. Finding the optimal type, dosage and timing of reagents as well as optimal physical variables, such as cell transfer time or temperature, is difficult and requires an enormous amount of testing.
To make this procedure more efficient and practical, a research team led by Genki Kanda of the RIKEN Institute in Japan set out to develop a self-contained experimental system that could determine the optimal conditions and The functional retina may develop pigment layers. Retinal pigment epithelial cells were chosen because degeneration of these cells is a common age-related disorder that renders people unable to see. More importantly, transplanted retinal pigment epithelial layers have already shown some clinical success.
For autonomous experiments to be successful, robots must repeatedly perform the same series of precise movements and manipulations, and artificial intelligence must be able to evaluate the results and prepare for the next experiment. The new system accomplishes these goals with a general-purpose humanoid robot called Maholo that is capable of high-precision biological experiments. Maholo is controlled by artificial intelligence software that uses a newly designed optimization algorithm to determine which parameters should be changed, and how they should be changed, so that the next round of experiments can lead to differentiation. efficiency can be improved.
In what would have taken human researchers more than two and a half years, a robotic system with artificial intelligence took only 185 days, and this translated into an initial differentiation rate efficiency of 50 percent to one of 90 percent, thanks to experimentation and performance by robots. Improvement work done.
Kanda and his colleagues uncovered technical details of their progress in the academic journal eLife, under the title “Robotic Discovery for Optimal Cell Culture in Regenerative Medicine.”