Over the past several decades there have been major developments in laboratory automation stemming from the emergence of systems capable of processing large numbers of samples in parallel [3-7]. Laboratory automation can range from simple automation, such as automated analysis instruments (i.e., Cell Lab Quanta® by Beckman Coulter, Inc.), to more-complex integrated systems, such as intelligent laboratory workstations (i.e., Biomek® 2000 Laboratory Automation Workstation by Beckman Coulter, Inc.).
The range of laboratory automation can be broken into two levels of user input: open-loop and closed-loop experimentation. In open-loop experimentation, an experiment or analysis is set up and run with no decisions made based on data from previous or ongoing experiments. More often than not, the experiment or analysis was only meant for serial evaluation, and decisions are not necessary. This allows for easy use of the instrument by scientists, but it restricts the utility of the instrument because few, if any, adaptive changes can be made to the experimentation. In contrast, in closed-loop experimentation, ongoing experiments are evaluated, and future experiments are pruned, altered, or spawned based on data from previous or ongoing experiments . Current integrated systems are moving in the direction of closed-loop experimentation, which is flexible and configurable and would allow the automated system to be a walk-away device. In return, this approach offers the prospect of increased productivity while reducing scientist intervention.
One of the more beneficial aspects of integrated systems stems from powerful software that allows scheduling of experimentation. Multiple sets of experiments can be implemented in parallel through the use of a scheduler. A simple scheduler offsets the start time of intact experimental plans and interleaves (in a comblike manner) the individual commands of the respective plans. The resulting schedule consists of a set of experimental plans with offset start times; in this manner the total duration of the set of parallel experimental plans is generally compressed by up to tenfold compared with that for serial implementation . More-complex schedulers exist that can order the experiments in a particular fashion, such as by user ranking or by shortest experiment duration first.
Automated experimentation instruments, created from the combination of computers with robotics, have been assembled for diverse applications ranging from high-throughput screening to library preparation to reaction optimization. Because there is a push toward more-intelligent automation, our efforts will concentrate on reaction optimization, which provides the most desirable benefits of intelligent design.
Was this article helpful?