Laboratory Automation – Industry Transformation or Dead End
Author Dr. Martin Pfister - Consultant
In April 2003 the Human Genome Project presented the complete version of the human genome about four years earlier than planned - but finished second. In the race between the public multinational consortium and Craig Venter’s Celera Genomics it was an automated ultra fast sequencer, which turned the balance in Celera’s favour.
Since the 1990’s, automation has affected the way drug discovery is handled and how diagnostic laboratories process samples in their daily routine. Ideally, laboratory automation in both the academic and the industrial setting should: · remove bottlenecks frequently caused by insufficient resources and manual handling · remove the need to process numerous samples with accurate reproducibility to provide good statistical data · remove the need for precise handling of very precious samples manually thus eliminating contamination · reduce costs · operate as fully automated “closed” systems (also called “walk away”) that are able to handle any sample but still being as flexible as possible.
How automation started - the “titration device”
The market for lab automation has emerged very recently. Basically, the history of lab automation parallels the development of modern drug discovery within the pharmaceutical industry. However, the trigger was the advent of a standardized format microplate. The microtiter plate, also commonly referred to as the microplate, was originally developed in the 1950's. One of the earliest patent applications for this type of device was granted in 1960 in the US. This patent describes a “titration device” that “…requires no manual operation of the parts thereof during titration, and which provides positive indication of the volume of reagent delivered from the pipette element thereof, without requiring the operator to read a meniscus.”
The 96-well microplate format was applied to scientific assays such as ELISA's in the 1970's, and has become a ubiquitous tool since. Today, the 384-well format is becoming the norm in high throughput screening by increasing the capacity of the original by a factor of four as well as offering less usage of sample, solvents, and reagents. With the 1536-well plate, the trend is being continued. While keeping that number within the same external dimension the resulting density stretches the ability of most of today’s available liquid handlers and other robotics.
For a laboratory with a throughput of about 200 plates per day, the usage of plates with different densities means:
From automated “high throughput” to “high content”
High throughput screening (HTS), commonly defined as the analysis of 10,000 samples per day, was being developed to search large numbers of potential drug candidates for activity against a specific disease. These potential drugs form compound libraries, which become the intellectual property of each drug company.
With the advent of new technologies such as microarrays, microfluidics, and “lab-on-a-chip”, not only the throughput but the way in which data is generated will change the face of drug discovery and clinical diagnostics within the next few years. Automated arrayed assay systems integrating multiple assay and detection methods will become major competitors to established clinical analyzers both for POC and clinical laboratory applications. Seth Pinsky, chief technology officer at MDL Information Systems has commented on this shift in complexity, “We were used to getting very simplistic results.” “Now we’re seeing more complex assays lending themselves to automation.” The industry has concentrated resources on HTS but now a new concept of high content screening (HCS) is emerging. This involves the similar usage of high productivity techniques but involves the automation of information-rich biological assays for the discovery and validation of new drugs. The new technical array platforms enable researchers to pursue more targets simultaneously while saving costs on labour and sample preparation as well as the use of even smaller volumes of reagents than necessary for a 1536-well plate. As a consequence, the industries will shift the focus on screening large numbers of compounds to the issue of what to do with the information scientists get.
Companies like deCode, headquartered in Reykjavik, Iceland, are already combining automated analysis methods with improved bioinformatics to resolve the pathobiology of yet unmet diseases within the shortest time frames. This process allows the identification of markers from huge databases that ultimately result in diagnostic tests.
Flexibility versus robustness?
The difference between science at the benchtop and that within the industrial setting is now being marked by the quality of automation used in those environments. Beckman Coulter Inc., headquartered in Fullerton, CA clearly dominates both – the workcell and the large scale and laboratory sectors.
From the benchtop viewpoint, the increasing trend is towards the implementation of modular automation - small workcells that can be rapidly configured, installed, brought into operation, and reconfigured when the need arises. Technical solutions are faster mechanisms and distributed robotics - that is, many simple, rather than one large robot. Thus, a workcell can be thought of as a small, automated solution that addresses tasks such as sample extraction, reading plates, or some other portion of an assay as a single device or integrated. “For me robotics means that you just integrate several different instruments into a work flow” a global business manager for automation at Qiagen GmbH says. Within that concept, the pipetting workstation – or liquid handlers - remain the cornerstone of lab automation and play a critical part in completing the automation process. With worldwide revenues of about $620 million and annual growth between 10 and 13%, automated liquid handlers are a large, dynamic and rapidly growing market. Within that market, diagnostics applications constitute the largest segment followed by genomics and proteomics. The advanced workcell may be capable of performing most or even all steps of any given assay – if the correct portfolio of automated equipment is chosen. Already over half of all clinical microbiology laboratories have installed automated instruments for detecting bacteria and for performing antibiotic susceptibility testing
From the industry side, stable platforms based on robust and reliable automation promise industrialised routine approaches and screening infrastructures in a large pharmaceutical business.
For a pharmaceutical company with about $810 million to spend on R&D for bringing one new product to the market, more value out of this investment must be the first thought. To this end, automated screening devices must rely more on robustness – that is the mean time between failure and accuracy over many thousands of cycles. In the pharmaceutical industry, the guarantee of successful screening is more important than the novelty of the approach. No production unit can bear the burden of constant innovation on its platforms. Therefore, the industry-wide drive to minimise screening costs must be matched by a willingness to develop and fund a reliable standardized automated infrastructure as demanded at the first conference on “Lab Automation Europe”, held in the autumn of 2003. Proper standardization that will allow vendor-independent modular configurations will assure the success of automated technologies.
Total automation – still a niche market?
The concept of the totally automated laboratory has not expanded beyond a few larger laboratories. In the haemostasis industry for example, the trend towards fully automated systems offering a wider selection of coagulation tests away from the electromechanical clot timers that have been in use for many years is irreversible. In principle however only the minority of labs is suitable for total laboratory automation (TLA).
In the USA for example only about 7% of the laboratories in the country are considered to be able to benefit from TLA
A 550-bed hospital and below is not suitable for TLA, unless they have a large outpatient business. As a rule of thumb, experts believe a laboratory should be performing at least 1.5 million tests per year before installing a TLA system. The potential market is somewhere between 300 and 500 in North America and a similar number in Western Europe.
Industry experts believe that a typical mid-size to large lab in the US processes 2500–3000 tests per day. With implemented automation using robotics, the lab can increase test volume by some 20%, reduce the sample turnaround times by 11%, and save about $100,000 in staff salaries. In many cases a TLA system will pay for itself in three years.
Thus in most of the laboratories using automation, the modular "workcell" concept has met with greater acceptance. The use of TLA in the clinical laboratory setting and the future direction that TLA may take in this setting is shown in the figure opposite.
Large-scale systems, while powerful, still suffer from some limitations as described in the table below.
|
What drives the automation market?
Major breakthroughs will – as in every market – drive the productivity and thus automation. Influences will come from the technical side as well as from diverse developments of the life science industry: Pharmaceutical companies desperately need an increase in effectiveness to stay cost- effective. The biotech market generates numerous demands fuelled by newly launched genomics and proteomics initiatives. They are likely to provide a great impetus to sustaining market growth in the future need for higher throughput and improved data handling. Another driving force is the ability to work with increasingly smaller quantities of material in biotech.
Robotics and automation benefit from increased processing power, memory growth, and increased communication bandwidth up to Gigabytes. This increase has spurred growth in enhanced data collection, data movement and sharing, as well as growth in an industry centred on remote smart sensors.
Outlook
Successful companies must be able to manage the split between bringing the cost down through less expensive reagents and technology and improving the robustness of the technology. What sets automation technology apart from so many other efficiency solutions are dramatic savings that it brings to the clinical laboratory. Expenses, directly related to biological and clinical diagnostics have already decreased over the past few years, despite the broader range of parameter detected and the increasingly sophisticated technologies used.
As new and innovative instruments with better features emerge, it will become incumbent on manufacturers to make the end user aware of potential applications. Laboratory automation can no longer be restricted to highly trained specialists - anyone working in the lab must be able to use them. With a correctly chosen portfolio, automated systems might be the silver bullet for effective, robust and “walk away” laboratory routines in both the benchtop and industrial environment.