SLAS

Automating LUO's: Strategy and scale

From LabAutopedia

Jump to: navigation, search
Invited-icon.jpgA LabAutopedia invited article


Authored by: Steven D. Hamilton, Sanitas Consulting

There are various system architectures available to automate a laboratory procedure. All are similar at a high-level, addressing the automation of one or more the three categories of Laboratory Unit Operations (LUO's): Sample Transport, Sample Processing and Data Collection & Handling. The mid and low level details of the various architectures can differ to a large extent, so an understanding of these details is essential for making good system decisions.  The LUO concept allows us to easily break down, understand and evaluate laboratory automation architectures.

Contents

Laboratory automation architectures

Device

We define a Device as: A laboratory instrument or tool that is capable of performing a single LUO. For instance, a motorized stirrer or a vortexer performs one LUO, mixing. A simple incubator or a hot plate performs one LUO, incubation.  Devices may be used in a stand-alone mode, with manual transport of samples and event execution.  They may also be incorporated into automated systems to perform their single LUO within the envelope and control of the system. 

Workstation

We define an Automated Workstation as: A laboratory device that is capable of performing a limited set of Laboratory Unit Operations (LUO's) (as few as two) in an automated mode, coordinated by a workstation controller.  Workstations are generally designed to fit into a typical laboratory environment, i.e. a laboratory bench.  Workstations can generally be purchased as "off-the-shelf" systems, with more complex workstations having a number of optional configurations.  Workstations are often reconfigurable within the scope of their defined LUO set, but fewer workstations are designed to expand or change the defined LUO set.    Using the LUO concept, we can examine the taxonomy of the workstation architecture.

  • Sample Transport: Samples or sample vessels are transported to/from the workstation either manually or by a general purpose transport device. Within the workstation environment, samples may or may not need to be transported.  Intra-workstation sample transport may be electromechanical or fluidic
  • Sample Processing: Performed by a collection of devices within the workstation envelope, coordinated by the workstation controller.   
  • Data Collection & Handling: The processed sample is subjected to a typical laboratory analysis, such as spectrophotometric or chromatographic analysis and the data are collected electronically.  This process may or may not be integrated into the workstation domain.   

Today, automated workstations have a very wide installed base.  Manual (sneakernet) transport of samples among a collection of automated workstations has proven to be an effective and flexible mode of automating a laboratory process.  Common automated workstations range from sophisticated electronic balances, multi-function microplate readers, automated liquid handling workstations, to specialized, multi-LUO sample processing workstations.  Microfluidic systems also generally fit the definition of a workstation.  Workstation controllers are often, but not always a PC, and most utillize some form of Graphical User Interface (GUI) and Method Based Programming.  More complex workstations may also offer a Scripting Language for programming functions outside the scope of the GUI/Methods Based Programming.  Most workstations also employ a Command Language, underlying the GUI/Methods Based Programming, to execute the base level electromechanical functions of the device.  Command Language programming is not always accessible to workstation users, and requires a reasonable level of programming expertise to utilize. 

Integrated System

We define an Integrated System as: A laboratory automation system consisting of multiple discrete devices or workstations, connected by one or more general-purpose automated transport systems.  Such systems are capable of performing many LUO's. They are custom configured for a specific application and are often reconfigurable or expandable, but most likely not in a totally plug-and-play manner. Thus reconfiguration requires specific system knowledge and expertise. Integrated systems are controlled by an overall system controller (i.e. a PC), and the incorporated workstations may be also be controlled by the same system controller or by their own dedicated system controller.  Integrated systems may require scheduling software to manage the workflow of samples among devices or workstations by the transport system(s).

  • Sample Transport: Samples or sample vessels are transported by a general purpose transport device. Samples or sample vessels are passed to/from various workstations or devices, each of which performs a specialized set of sample processing LUO's.
  • Sample Processing: Performed by simple devices or specialized workstations, serviced by the general-purpose transport device.  Workstations range in complexity, with examples such as electronic balances, liquid dispensers, centrifuges, capping/uncapping stations, filtration stations and complex mixers.
  • Data Collection & Handling: The processed sample is subjected to a typical laboratory analysis, such as spectrophotometric or chromatographic analysis and the data are collected electronically, This process may or may not be integrated into the workstation domain. 

General purpose transport approaches for integrated systems may take several forms, such as robots, linear transport devices (conveyors), positioning stages, fluidic channels and combinations thereof.  Integrated systems may range from laboratory-bench scale to room size, but the fundamental architecture remains the same.  Large systems may combine one or more transport systems and may have multiple branches of transport paths. 

Automated Workcell
It could be argued that some of the most complex workstations, have, in fact, become more like integrated systems.  It could also be argued that some of the simpler integrated systems are more like workstations.  Definitions must evolve as technology evolves, thus we now define a sub-category of Integrated System, the Automated Workcell:  A special case of integrated system,  preconfigured, often available commercially off-the-shelf as a standard system for a given type of or class of sample processing. As compared to custom, one-off integrated systems, workcells tend to be more compact because of more focused and on-going engineering efforts, and may not be as easily reconfigured outside of the pre-defined process focus. Otherwise, workcells are integrated systems, e.g. employing multiple devices and workstations linked by an general-purpose sample transport system. Often an automated liquid handling workstation is a key component of a workcell.

Random access vs. Linear

Automated systems can be designed such that they offer random access or linear sample processing.

Random access model

  • System devices and/or workstations are used for multiple occurrences of a given LUO(s) within the automated procedure.  For instance, the same microplate washer may be used for two or more occurances of washing the same sample plate.  The system may be programmed to move a given sample to a given device/workstation at any time in a process. 
  • Sample flow is multi-directional.  The transport mechanism must be capable of moving sample vessels to/from the same workstation/device multiple times.  Robotic arms are multi directional.  Conveyor belts, although linear in nature, can be designed to move in both forward and reverse directions, and so could return a sample vessel to a workstation/device that it had visited previously.  Fluidic channels can also flow bi-directionally, but doing so may complicate laminar flow or diffusion conditions.  
  • The transport mechanism is often rate-limiting.  
  • System scheduling software is usually necessary to optimize and control the multiplexed access of sample vessels to devices/workstations.  
  • Random access systems generally have slower throughput than linear systems. 

Linear model

  • System devices and/or workstations are used for only a single occurrence of a given LUO within the automated procedure. For instance, a given microplate washer may be used for only one occurance of washing the same sample plate.  If the procedure calls for multiple plate washing steps, the system must contain multiple plate washers. The system moves a given sample to a given device/workstation only once in a process.
  • Sample flow is uni-directional.  Once a sample visits a given device/workstation, it will never return to that same workstation.  
  • Devices/workstations are rate limiting.  Specifically, the device/workstation with the longest sample processing time sets the pace for the entire sample flow.  
  • System scheduling software is usually not required.
  • Linear systems generally have higher throughput than random access systems.   

Parallel approaches

There are cases when individual systems do not have sufficient capacity or throughput to meet requirements.  Rather than overly compliate the engineering of a system in an attempt to boost capacity or throughput, often the simplest and most effective approach is to employ multiple systems to provide "parallel processing".  This parallel processor strategy has been successfully used in the microprocessor industry to increase computing power.  Perhaps the most public example of this approach toward laboratory automation was the Celera Genomics decision to purchase approximately 100 commercial automated liquid handling workstations to support the scale-up of their Human Genome mapping effort, rather than invest the time in developing custom, higher-throughput systems.  Employing multiple, parallel electromechanical automation systems can be quite expensive, whereas the same approach for fluidic systems can be quite economical.  Using semiconductor fabrication technology, it is quite possible to create multiple, parallel microfluidic systems on the same substrate.  Many examples have been reported[1][2][3].  Optical detection for such devices is more complex, but has also been accomplished.

There are other, economical ways to accomplish parallel processing with electromechanical automation via employing parallel or arrayed sample vessels.  Using this approach, electromechanical devices are designed to accomplish sample transport, sample processing and measurement on multiple samples simultaneously.  The most common example of this approach is the use of microtiter plates, or microplates.  A microplate is a plastic-molded array of small sample vessels, or "wells".  These exist in arrays of 96, 384, 1536 wells, plus other custom densities.  The arrays of samples contained in a microplate are transported, processed and measured as a group, thus increasing system throughput.  Microarray technology is simply a more miniaturized and much higher density approach, with samples placed on a flat substrate.  There are even cases where microarray grids are printed on the bottom of microplate wells, to provide multiple different experimental environments for an array of microarray grids.

Another interesting example of a parallel electromechanical approach combines the array and linear transport technology.  The Tape Array[4] (Global Arrays) employs sample well arrays embossed continuously into a sprocket-driven plastic tape.  In this way, the sample vessel and sample transport technology are essentially combined in the form of the tape.  The company sells instrumentation to address liquid handling, array sealing, and assay reading for this product. 

Examples of parallel sample processing arrays
Microplates: 96, 394, 1536 well density
Microarray Grid
Array Tape

Related articles

References

  1. Brenan, Colin J.; Morrison, Tom; Stone, Kristine; Heitner, Tara; Katz, Arrin; Kanigan, Tanya S.; Hess, Robert; Kwon, Soek-Jooh; Jung, Heung-Chae; Pan, Jae-Gu, Massively parallel microfluidics platform for storage and ultra-high-throughput screening, Proc. SPIE Vol. 4626, p. 560-569, Biomedical Nanotechnology Architectures and Applications, 2002
  2. Yairi, M. and Richtera, C., Massively parallel microfluidic pump, Sensors and Actuators A: Physical, 2007, Volume 137, 350-356
  3. Medintz, I., Paegel, B. and Mathies, R., Microfabricated capillary array electrophoresis DNA analysis systems, Journal of Chromatography A, 2001, Volume 924, 265-270
  4. Elaine, M., Array Tape for Miniaturized Genotyping, Genetic Engineering & Biotechnology News 2007-06-15
Click [+] for other articles on  The Market Place for Lab Automation & Screening  Robots-Automated Systems Integrated Systems & Services