Despite protestations to the contrary, automation is not something to undertake casually. Due to the dizzying array of possibilities and workflows, the technology has not yet achieved the simplicity of consumer products.
Given the diversity of analytical samples, the first few steps of sample preparation have little in common. Consider how one might acquire and initially process an orange, oil sludge, soil, or water. The first few steps might consist of filtering, grinding, adding stabilizers, sonicating, or a hundred other things. Convergence in prep operations and techniques occurs as the sample gets closer to the actual analysis. At this stage, diluting, dispensing, reconstituting, adding reagents, chemical derivatization, and mechanical operations such as heating, cooling, mixing, and extracting become more standardized.
One area that remains challenging for automation is primary sample handling: acquiring samples and getting them into a suitable format for further study. A geologist or an environmental scientist must still travel to the field to collect rock or water samples and get them ready for lab work. Some primary steps, such as drawing blood or excising tumor tissue and rendering it into a plate or tube from which it is accessible to the automation system, involve a very high level of skill. “While these examples require a good deal of human intervention, there are opportunities to address those types of workflows as well,” notes Jeremy Lambert, director, automation and liquid handling at PerkinElmer (Hopkinton, MA).
A typical sample workflow may involve collecting the sample, subjecting it to the reagents and purification, transferring it to some form of labware, applying physical/ mechanical conditions (heating, cooling, shaking, amplifying), and introducing it into the analysis system. “A lab technician can perform all these operations by hand,” notes Mehul Vora of Beckman Coulter, which just introduced the latest product on its Biomek platform, the Biomek 4000 Laboratory Automation Workstation. “But the scientist is after only the data.” Nothing that occurs before that DNA analysis or spectroscopy reading adds value to the scientific exercise, Vora explains. “If scientists could wave a magic wand and have the sample ready for analysis, they would do that. That’s not the real world, but the relevant question is what is the best way to get there?”
Limitations of sample prep systems vary by instrument type and sample. Most are related to the fact that samples tend to be fluids—but not all liquids are created equal. Viscous samples require more work, time, and care than standard samples do. “Many viscous samples are difficult to filter,” observes BioTek product manager Jason Greene. “You may have to increase the vacuum, pull on the sample longer, or watch it more closely to get through that step.”
For drug assays, the predominant solvent, dimethyl sulfoxide (DMSO), presents assaying problems due to its hygroscopicity. Solutions rapidly equilibrate in laboratory air to 70 percent DMSO and 30 percent water, which can interfere with many assays or provide erroneously low concentrations. Companies looking for very accurate inhibition constants for drug candidates need to take this into account.
From a concentration standpoint the inverse of hydration is evaporation, which is always an issue with very low-volume, water-based assays. Greene hears about this problem “a lot, which is why many customers have backed away from ultralow-volume 1536-well plates and are beginning to readopt 384- and 96-well formats.”
Finally, Greene warns lab workers to know the limitations of everything they work with, including complex sample prep systems. “It’s funny how diligent scientists can be in testing everything except instruments. They don’t realize the wide variability that can occur over time. It takes way longer to troubleshoot bad results than to perform due diligence up front.”
Vendors of prep equipment and systems, reagents, kits, and labware strive mightily to accommodate the need to automate routine lab tasks, but what are the limits? We know that an automated liquid handler cannot grind tissue or dispense many liters of solvent at once. Yet they have the bases covered for most preparative workflows in the life sciences.
“When we ask our experts, they say anything related to life science workflows is possible,” notes Mehul Vora. “Give me the assay and I can automate it, provided it fits into a tube or a microtiter plate.” What about an assay that does not fit the standard ANSI/SLAS (formerly ANSI/SBS) format or was previously not automated or is perhaps too complex and operator-centric for standard automation tools? With the input of assay vendors, engineers from Beckman’s Integrated Solutions division can design labware that enables automation of many customer workflows.
For labs unfamiliar with automation, particularly for sample prep, vendors such as BioTek Instruments serve a vital role in dispelling doubts and opening up vistas. “We know a lot of super high-end systems are out there,” Greene observes. “What we offer is that same general capability in a smaller, more affordable format. Like many automation firms, BioTek reaches out to equipment and reagent companies to combine their analytical capabilities with BioTek automation systems. Examples include an off-the-shelf vacuum manifold that required a special labware holder and an oligonucleotide sequencing reaction cleanup kit, which was automated on a BioTek platform. Both additions resulted from a collaboration with Millipore. BioTek’s other automation partners reads like a Who’s Who of laboratory equipment: Caliper, Hamilton, Agilent, Tecan, Beckman Coulter, and others. These efforts are “very well received” by reagent, kit, and labware companies as well as by customers, Greene says.
Despite operating at what one might call the entry level of automation in terms of throughput, BioTek is integrated within the automation market, selling detectors, robots, liquid handlers, and data systems. Of the latter, Greene believes that the software and interface should match the ease of use of the automation system itself. “Users don’t want to spend days writing a liquid handling protocol.” High-end systems can offer significantly more capability but cannot always boast the same user-friendliness.
Care and maintenance of automated sample preparation systems are difficult topics because so many types of “sample prep” exist. Limited to liquid handling systems as we are, one finds the usual issues of replacing valves and washers and maintaining cleanliness. Rinsing is critical after a run on a liquid handler. With pipetting tools, O-rings may require replacement every few months, depending on frequency of use and type of liquid. Simple tests determine whether a pipette seal has been compromised.
A common issue with automated microplate washers, for example, is buildup of proteins and salts, which clogs equipment. The most effective way to remove this buildup is through ultrasonic cleaning. The manifold is typically removed from the instrument and placed in an ultrasonic bath for cleaning. BioTek offers integrated ultrasonic cleaning so that the manifold can remain in place and the cleaning can be preprogrammed and run without user intervention.
In the end, failure to understand what manufacturers recommend for routine care is one of the biggest maintenance issues, says Greene. “There are things with liquid handlers that you should be doing daily, weekly, monthly. Given the high turnover in some labs, this doesn’t always happen. When things are let go for too long, you wind up with cumulative errors—you reach a point where your data is no longer within acceptable tolerances. Care and maintenance don’t take a lot of time, but they need to be done based on the manufacturer’s recommendations and the lab’s workflow priorities.”