U.K. researchers create robot biotech scientist

Time for researchers to put their feet up? A team of U.K. scientists has developed a robot scientist to do their job for them.

Or perhaps not quite. The robot, and the computer system with which it works, have been developed to help generate hypotheses about the function of particular genes on baker's yeast, and then carry out experiments to test them. So far the problems it is solving are fairly simple, but the technology could make a profound difference to biotechnology research in the future, according to Stephen Muggleton, a professor in the department of computing at Imperial College London, one of the team members working on the robot project.

The robot, which is based at the University of Wales in Aberystwyth, and its computer system are never going to replace human researchers, but will make it possible for them to work more quickly, Muggleton said.

"In the area of bioinformatics, the amount of data being produced is increasing at an exponential rate. There are not enough people to look at all the data. (The robot) is not going to lead to redundancies, but instead will allow people to deal with the work," he said.

The robot itself is a simple machine with no onboard intelligence. It places different strains of yeast, each of which have one gene missing, known as "knockout" strains, into a 96-well plate and observes how they grow on different chemical substrates. From this it is possible to establish possible functions for the missing genes, the scientists said in a statement.

Muggleton designed a closed-loop machine-learning program that takes the information about the yeast growth from the robot and creates hypotheses from that data. The program then identifies the most likely hypotheses and decides which further experiments should be performed. Those choices are fed back to the robot, which will then perform them.

The machine learning program itself runs on a standard PC, though larger computer systems would be needed for more complex research, Muggleton said.

The approach is a very general one that could be applied to lots of research tasks, Muggleton said. Drugs companies in particular may be interested because the amount of time and money spent on analyzing different molecules for a new drug could be reduced, he said.

Imperial College Innovations, the college's commercialization arm, is looking at how the product can be sold, perhaps within two years, Muggleton said.

"Initially we will try to work closely with the pharmaceutical companies, on products designed specifically for them. It will be a while before there's an actual product to sell," he said.