Pipettes to Robots
Systems exist for nearly every lab, workflow, and throughput level, even for many labs that believe automation is too complex or too expensive.
High-end liquid handlers serve high-throughput workflows at large companies, but all automated systems share one characteristic. “They all replace tedious manual operations, which is where most errors occur,” says Ian Shuttler, head of strategy and portfolio management at Tecan (Männedorf, Switzerland).
Full-blown robotic liquid handling systems are formidable, integrated systems with steep learning curves, but complexity is somewhat mitigated through improved interfaces. Tecan, for example, has introduced a simple touch-screen interface for working through methods with the help of graphics, text, and application wizards.
At the mid-range level of complexity, vendors are taking advantage of their entry- or mid-level automation products to create application-specific systems that may be reconfigured down the road as workflow demands change. This product development strategy requires balancing immediate needs with future-focused flexibility. “As customers gain experience, the systems should grow with them, so robotics-enabled liquid handing systems remain relevant,” Shuttler adds.
What’s unique compared with ten years ago is that the market is addressing users who lack the automation and coding experience demanded by larger automation systems. “Everyone needs some way to move liquids around, so the key is to make liquid handling more accessible and easier to integrate with hardware,” says Tara Jones-Roe, marketing manager at Beckman Coulter (Indianapolis, IN). Most labs that purchase the company’s entry-level Biomek 4000 liquid handler, for example, are first-time automation buyers.
The original, big-money driver for lab automation was high-throughput screening of drug candidates. While that market still exists, the “numbers game” has reduced from millions or hundreds of thousands of compounds to just hundreds or thousands.
Paradoxically, the market for liquid handling systems has increased as screening numbers dropped and labs increasingly look for greater consistency, says Michael Beier, product manager at Integra Biosciences (Zizers, Switzerland), which specializes in liquid handlers that straddle the complexity continuum between handheld pipettes and fully automated systems. “Less hands-on time allows workers to concentrate on other tasks, but also improves ergonomics, especially for veteran lab workers who are prone to repetitive stress injuries,” he explains.
One result of the automation “craze” is that labs that would not have considered automation a few years ago are now buying into it, particularly for benchtop, semi-automated liquid handling systems. “People are looking for instruments that are affordable, that they can operate easily, and whose operation doesn’t require dedicated personal,” Beier says. The Viaflo 96 handheld 96-channel pipette, for example, represents an interesting crossover product between handheld and low-level automated liquid handlers. “Anyone can use it, and it’s affordable,” Beier adds.
The question of application specificity versus flexibility is one that customers and vendors continue to grapple with. “There are two sides to that coin for sure,” says Jason Greene, senior product manager at BioTek Instruments (Winooski, VT). BioTek’s modular approach, based on components of automation as opposed to large robotic systems, works well for labs whose needs change frequently—provided they possess the expertise to take advantage of that flexibility, or the funds to call someone in to make the adjustments.
In other situations, well-funded customers who do not anticipate significant assay changes might consider large, high-end automated workstations. “While that’s a fine option for other people who need more flexibility, the big liquid handlers are perhaps a bit too rigid for them in terms of capabilities. That’s where the smaller, modular approach may be more appropriate,” Greene adds.
Beckman Coulter is in a similar position technologically. As a member of the Danaher family of life science companies, Beckman is able to exploit sample preparation expertise from its sister companies that extends into cellular, genomic, and proteomic applications. And like many automation companies, it regularly collaborates with other automation companies to deliver customized products. Lab automation in general, and liquid handling in particular, could not progress as far as it has without these partnerships.
Role of mid-range systems
The trend toward miniaturization and personalization has fueled demand for mid-range liquid handlers that are more sophisticated than automated manual pipettors but lack the complexity and capabilities of large robotic systems. Joby Jenkins, product manager for liquid handling products at TTP Labtech (Cambridge, UK), calls such systems “dedicated benchtop, low-volume pipetting instruments.”
TTP Labtech’s mosquito line of liquid handlers, for example, lacks many of the attributes of full-scale robotics from Tecan or Hamilton. As the name implies, these liquid handlers are smaller and deliver volumes in the low to mid-range, from 25 nanoliters up to 5 microliters.
Compared with many large robotic systems, mosquito is viewed as low to medium throughput. “However, for certain applications, such as protein crystallography, being able to set up nanoliter volumes in a ninety-six-well plate in less than two minutes represents a significant degree of miniaturization and increase in throughput,” Jenkins says.
Moreover, the systems integrate well with robotics. Users can employ a plate handler and stacker, or can aim for even bigger things. TTP Labtech has partnered with Tecan, for example, to integrate a mosquito with Tecan robotics and software, which extends the capabilities of both systems. “The user’s experience is of using a large Tecan system that now can pipette lower volumes. The whole integration is controlled by the one software,” Jenkins says.
Two related trends in automated liquid handling have emerged from customers’ desire for personalization. The first involves less reliance on core automation facilities; the second is the need for systems dedicated to one task.
“Today’s scientists are looking for their own personal benchtop pipetting systems that do one thing very well,” explains Bobby Chavli, associate director of marketing for Hamilton Robotics (Reno, NV). “This represents a big change.”
One driver for benchtop robotics is the realization that automation’s benefits transcend high-throughput capabilities. Laboratories increasingly appreciate reproducibility, consistency across operators and labs, error reduction, and walkaway time.
Related to personalization is the desire for off-the-shelf “solutions”—systems customized for specific assays using kits from a single reagent vendor. For example, Hamilton’s Nimbus next-gen sequencing workstation works specifically with Roche kits, whereas the NGS workstation is configured for the New England Biolab and Illumina chemistries.
Hamilton still sells general-purpose robotics platforms such as Star, Vantage, and Nimbus, whose prices range from about $100,000 to $2 million. Application-specific modules are built on these to a high degree of specificity. And core automation facilities still operate for complex workflows and high-throughput projects, but the benchtop norm has been redefined toward greater capability and, simultaneously, user-friendliness.
Whereas five or ten years ago customers may have spent weeks or months developing automated methods for these assays, today’s user is more focused on the application itself and on hands-free time that may be spent more productively. “Now, if they use Illumina reagents, they expect an automated workstation for those experiments,” Chavli says.
Care and maintenance
Calibration and maintenance have always been sticking points with instrument users. This is true with liquid handlers as well, whether dealing with single-channel manual pipettes or dispensing heads for large systems. “Many customers would be happy never to calibrate their pipettes,” says Melinda Sheehan, product manager for liquid handling at Eppendorf (Hauppauge, NY).
Labs have many convenient options for pipette calibration. Kits are available for in-house gravimetric calibration by technicians or end users. For example, Artel (Westbrook, ME) sells a calibration workstation, the PCS®, for do-it-yourselfers.
Lab managers often turn for calibration to original vendors, who provide on-site calibration services for large groups or departments or service via standard delivery services. Some manufacturers even provide loaner pipettes when turnaround is expected to be lengthy. Numerous third-party service organizations will service pipettes at their shops or at the customer’s site as well.
Despite these choices, many users simply will not part with their favorite pipette for a week or have a liquid handler sit idle for a similar length of time.
“Many customers come up to us at trade shows and ask about pipette calibration,” Sheehan says. “By the time they’re asking that question, they probably have a drawer full of broken or out-of-spec pipettes.”
Eppendorf has eliminated some of the burden of calibration record-keeping by installing radiofrequency identification (RFID) chips into all its manual and electronic pipettes. RFID stores more information than most pipette operators require to remain in compliance—for example, serial number, date and type of service, and usage. Eppendorf has used the chips “since before we even had a plan regarding what we would do with them,” according to Sheehan.
Electronic tagging practically eliminates the need to collect and store paper calibration certificates, which Sheehan describes as passé, saying, “Certificates get lost. In some cases the upper and lower parts of instruments get mixed up, and nobody knows which reassembled pipette the certificate belongs to.”
Eppendorf has recently introduced TrackIT software, which uses a USB-style scanner to read and store calibration information. TrackIT allows users to enter additional information as per company policy or regulations. Eppendorf has similar tracking systems for single- or multichannel pipetting heads used in automated liquid handlers.
Similarly benchtop systems, like Integra’s Viaflo 96, demand little upkeep. “The only part that requires service is the pipetting head,” Beier says. “The mechanics are maintenance-free.”
Calibration interval depends on the user’s established standard operating procedures, but usually involve sixor 12-month cycles. Robotic systems, by contrast, require service contracts under greater control of vendors, Beier explains, saying, “Mechanical parts need service.”
The final word on maintenance involves not instrumentation, but lab workers. “Users must continually be trained on how to use the latest good laboratory pipetting techniques, use pipetting products that offer the best ergonomics and performance, and regularly maintain their instruments with a certified service provider,” says Raymond Mercier, business director of liquid transfer at Thermo Fisher Scientific (Waltham, MA).
Pre-purchase considerations are as varied as liquid handlers themselves. Purchasing a handheld pipette, for example, is a low-budget personal choice, while a fully robotic liquid handling system is a capital expenditure.
For pipettes, Mercier suggests:
- Find a comfortable, ergonomic fit to prevent poor pipetting technique, which affects accuracy and precision.
- Look for pipettes and tips that were designed together.
- Don’t skimp on consumables: consistency in pipette tips reduces fit variability and benefits overall performance.
“And be sure to select a vendor with robust global manufacturing capabilities to ensure uninterrupted pipette tip supply,” Mercier adds.
Semi-automated systems are simpler than robotics and spare end users from extensive training. Integra’s Viaflo and Assist semi-automated systems, for example, are true walk-up instruments. “If you can operate a pipette, you can operate these instruments,” Beier says. “Lab managers typically believe that if manual pipettes are not enough, they have to splurge on a robotic system. That’s untrue. They should consider options in between, which may be more economical and, for many workflows, more efficient.”
It all comes down to understanding workflows, Beier adds. “Do they want to ramp up a little or a lot? Without fully automating, a ninety-six-channel pipette improves throughput significantly compared with a conventional handheld pipette.”
Managers need also to consider how many people will be using the instrument. If it’s a dedicated system, with a single task, with preprogrammed methods, and if walkaway time is valued, then a robotic system is appropriate.
“It sounds obvious, but the most considerations when purchasing automation are the requirements of your day-to-day workflow,” Greene says. Workflow includes individual process steps as well as specific liquid handling steps and bottlenecks based on anticipated throughput. Some processes or steps might work well with existing equipment, while it may make sense to automate other operations. Anticipated benefits range from increased walkaway time, accuracy, and precision to the ability to introduce uniformity to operations prone to human error. Lab managers might, for example, notice variations from technician to technician or at certain times of day when other things are occurring in the laboratory.
Other common considerations are the throughputs, batch sizes, and types or viscosities of reagents a lab works with. Purchasers should consider whether they purchase reagents or make them up in the lab, in what quantities, and the value of consistency.
Greene believes that dedicated, application-specific customization is a smart business strategy from the perspective of product differentiation. “If [customization is] done right, vendors can rest assured that they’ve done their homework,” Greene tells Lab Manager. “But we know from experience that few individuals purchase automation in that manner because they know that colleagues will want to use it. And if it goes into a core facility, a lot of people will be sharing it. As soon as the system enters the lab, they’ll forget the needs that were enunciated during the demo and sales process. That’s why you need, at least at some level, an open design.”
In other environments, when budgets allow, throughput is high, and/or assay requirements are relatively fixed, a large automated workstation can make sense. While this is a suitable choice for labs that don’t need total flexibility, higher-end, fully automated liquid handlers can be too rigid for many. This is where smaller, modular approaches can be more appropriate.
For Jones-Roe of Beckman Coulter, purchase decisions should be based on trusted partnerships with vendors, scientific expertise, product knowledge, and a high level of support in hardware, software, and applications writing. “Whether you’re spending $50,000 or $1 million, can an automation platform grow with your lab? Is it versatile or locked down? Can you add or remove components?” she asks.
Avoid the cul-de-sac
Despite the trend toward application-specific automation, users should be wary of entering an automation cul-de-sac. For example, Tecan has learned through customer surveys that many closed, dedicated liquid handling systems eventually go idle. Determining the most appropriate mix of current utility and future flexibility is therefore a critical part of early discussions with automation vendors.
Shuttler refers to system acquisition as a “consultative sale, where the salesman acts as an automation consultant; they know the systems’ capabilities and can recommend one based on current and anticipated future needs.”
Chavli notes that while customers are looking for “solutions,” liquid handlers can easily become expensive paperweights if they do not possess the versatility demanded by changing workflows. “Perhaps you purchased a liquid handler for PCR. Great. What happens when your lab is no longer performing PCR? Obviously, you want built-in flexibility, even though specific tasks are the reason for purchasing the system,” he says.
At some level, purchasers should seek systems that, with the assistance of the vendor or in-house automation experts, may be reconfigured for future projects. “In addition to the performance, precision, and accuracy required by current workflows, customers need to take a long-term view,” Chavli advises. “These are expensive systems that most users hold on to for many years.”
Related are the anticipation and realization of value across an automated system’s lifetime: the enhancements in throughput, quality, and time savings the system adds to your laboratory and results. Vendors of some systems claim, for example, to dispense 1 microliter volumes that are common in modern low-volume assays. Chavli questions such assertions unless the vendor can demonstrate that dispensers compensate for a fluid’s viscosity and density: Think concentrated buffer versus glycerol or oil.
Lab managers considering the purchase of an automated liquid handler must consider current needs while remaining open to future workflow demands. Labs should question vendors about upgradability and interoperability (particularly with respect to components already in the lab), assess available literature, and whenever possible put systems through their paces.