The three-day meeting was organized by Tony Beugelsdijk, a chemist specializing in laboratory robotics at Los Alamos National Laboratories. Many microbiologists tend to consider robotics an all-or-nothing proposition, he said, when in reality reprogrammable, multi-use robots and forms of single-task automation can be used together to speed up the sequencing process.
"There's a lack of appreciation for what automation can do at this point," Beugelsdijk said after he conducted a cross-country search on the potential applications of robotics. "[Biologists] thought robots were very good at spray-painting cars." To those who were disappointed by what they saw at the workshop, he pointed out that the application of automation to molecular biology is still "a very new field."
"There's been an acceptance problem," he added, based on the scientists' fear of losing personal recognition for publishing data compiled by robots. "If we accomplished anything, we planted some seeds in people's minds about the potential that exists."
Some skeptics remained, however. "The robotics were horribly primitive," observed Harvard's Walter Gilbert of those he saw displayed. Robotics were being used to duplicate the manual process, he complained, when the emphasis should have been on modifying the process so it could be done more easily by machines.
Gilbert supports the use of automation in a gene sequencing project. Like Beugelsdijk, he predicted that "the process will become more and more mechanized until it's done by technicians and not scientists." But Gilbert and others question the role robotics will play in automation. "The emphasis on the multipurpose gets in the way of doing any given task," he said.
The robot Riedel presented at the workshop—which is used in a preliminary stage of sequencing to isolate DNA—made use of a change from the manual procedure. It uses reagents whose chemistry has been modified for storage at room temperature. The manual procedure requires refrigeration of the reagents, a step that would have complicated programming of the robot.
The robot isolates plasma DNA from bacteria, handling 96 samples in 30 hours. Although this rate is slower than that performed by manual labor, Riedel said "the machine does as good a job [as humans] and frees up people from a tedious task."
Genetics Institute developed the robot as part of its research into developing proteins for use as therapeutic agents. The pharmaceutical company has not decided whether to market it, Riedel said.
Although not a robot, a DNA sequencer developed by Applied Biosystems Inc. of Foster City, Calif., created quite a stir at the workshop. "It's very exciting," said Riedel. "I think there's a lot of other applications for it."
Technology for the sequencer, which is to be tested shortly at six sites, was developed by Leroy Hood at his California Institute of Technology laboratory. It will be sold for $89,000, said product specialist Ann Wan. The machine is expected to perform at up to twice the rate of manual sequencers and to achieve accuracies of 95 percent or better.
Riedel predicted that some steps in the sequencing process will continue to be done manually. He also expects changes in biochemistry to eliminate the need for some kinds of automation.
Beugelsdijk said there is a pool of talent available at Los Alamos to advance the process, adding that the Department of Energy has a longstanding interest in sequencing for purposes of studying the effects of radiation. The laboratory's GenBank has about 10 million sequences currently on file.
The administration's budget for fiscal 1988 requests $11.5 million for the sequencing project and the GenBank. A department spokesman said some of this money might be made available for development of robotics.