Ross D. King from the University of Wales, Aberystwyth, and his colleagues designed their Robot Scientist using existing technology and two new software programs that they wrote. The team then assigned the robot the task of determining the function of specific genes in the well-understood yeast Saccharomyces cerevisiae (commonly known as brewer's yeast). Robot Scientist, armed with preloaded information about yeast biochemistry and biological pathways, first generated hypotheses regarding possible functions and then ran a variety of experiments. When the real scientists compared the results obtained by their robot student to those achieved by actual graduate students, they didn't see any significant differences. And because Robot Scientist ran fewer experiments, its overall costs were lower than those of its human counterparts.
"This research is very exciting as we have given the robot--under our supervision--the ability to design the experiments and interpret the data for us," King says. "There is increasing need for automation in the biological sciences, and although the problems we set for the robot were relatively simple, we have shown that it could be used to help solve real-world problems." Of course, the team notes that Robot Scientists will never fully replace grad students¿or their professors¿but they could provide human scientists with ample time to "make the high-level creative leaps at which they excel."