Alex Michel Ph.D., Principal Research Associate, Genzyme Corp., Waltham, MA
Luigi Francesco Covelli, Microbiologist, Medimmune, Mountain View, CA
Hugo M. Oliveira Ph.D., University of Porto, Porto, Portugal
Hardik B. Shah, Contract Chemist, Procter & Gamble, Mason, OH
Alex Nisthal Ph.D., Postdoctoral Fellow, Genzyme Corp., California Institute of Technology, Pasadena, CA
Q: What types of automated liquid handling instrumentation do you use?
A: Alex Michel: We use a variety of automated liquid handling systems in our lab. Primarily we use Tecan EVO 200s for the majority of our workload, but we utilize Tomtec and PerkinElmer ALHs in our weekly routine as well. Almost all of the systems are “stand alone,” and the assay plates we generate from those systems are almost all used in some type of LC-MS analysis.
Luigi Covelli: The majority of our liquid handling [instruments] come from the Hamilton Company. We employ Hamilton STAR robotics to run the assay setup for our process and then batch-process our assays onto smaller, point-of-use liquid handlers. These instruments include Thermo Fisher Scientific’s Multidrop with RapidStack plate handlers, and MDS Aquamax 2000/4000 plate washers with Stackmax plate handlers. For PCR-type applications, we use liquid handling robotics from Qiagen, specifically the QIAsymphony, M-48, Corbett CAS-1200, and QIAgility.
Hugo Oliveira: I'm a user of flow injection analysis and related techniques. In my case, the main goal of ALH is to ensure automatic and miniaturized sample preparation (analyte enrichment by solid phase extraction, or SPE) for liquid chromatographic analysis. The flow analysis manifold is coupled with the LC instrument for an automatic and integrated analytical protocol.
Hardik Shah: I work with the Rainin Liquidator and the Janus Multi- PROBE. We use the latter to transfer liquids from Bio Plas tubes [BIO PLAS, Inc.] to 96-well mass plates, typically at volumes from 20 to 1,000 microliters. The liquid handlers are not connected to any other instruments.
Alex Nisthal: We exclusively use a Tecan Freedom EVO system that has a lot of integrated bells and whistles, like a PCR machine, shaker, microplate reader, and vacuum unit. The Freedom EVO is contained within a biosafety enclosure, with waste lines and data cables leading out of the hood.
Q: Describe your “typical” workflows, and types of samples and experiments you run.
A: Alex Michel: Working in a DMPK department, our typical workflow centers around chemical property analysis, in vitro ADME studies, and sample processing for bioanalysis. We typically will use the systems to process samples for purity analysis, solubility assessment, and LogD determination. Acceptable compounds are then assayed for microsomal metabolic stability liabilities, CYP inhibition, and cell-based permeability, all of which are automated on our Tecan systems. Further down the line, we’ll also handle samples for bioanalysis, plasma-protein binding determination, efflux potential, hepatocyte metabolic stability, CYP induction, etc. Again, all of those assays are either fully automated on our ALHs, or at least partially automated on those same systems.
Luigi Covelli: Our company tests samples from clinical trials, including whole blood, serum, plasma, nasal wash, and cell culture supernatant. We run a range of assays on these samples, for example hemagglutination inhibition, neuraminidase inhibition, microneutralization, viral genotyping and subtyping, fluorescent focus potency, TCID50, and sample preparation. Typical automated assay workflows include sample ID verification, sample addition, serial dilution, transfer, and plate washing. Our use of automation is similar to a production line. Instead of automating the assay from start to finish, we use batch processing where specific steps are automated. With the implementation of our batch process automation, we’ve seen our throughput increase tenfold, and find this method suitable for our purposes.
Hugo Oliveira: My typical application comprises automatic SPE at microscale—either permanent or renewable sorbent approaches are possible—for analyte enrichment and matrix removal prior to LC injection. My work has been devoted to method development that comprised applications in the analysis of environmental samples, mostly water, and foodstuffs.
Hardik Shah: Approximately 90 percent of the time we use ALH to transfer liquid to 96-well plates for mass spectrometric biomarker assays. We run between three and four hundred such samples per day. Sample matrices typically consist of both aqueous and organic solvents, but occasionally human plasma. After transferring aliquots of samples to the 96-well plate, the samples are run on LC-MS-MS. The standards and QC results are quite consistent within and between batches.
Alex Nisthal: Typical workflows are operations that support our protein-engineering lab and another structural biology lab on campus. These include parallel site-directed mutagenesis, gene assembly, and protein purification. Assays include those measuring protein stability, HIV neutralization, and ELISA.
Q: What factors should purchasers of ALH instrumentation consider?
A: Alex Michel: Reliability, flexibility, ease of use, and redundancy all factored into our decisions. We wanted systems that we could count on, and that would also allow us to add onto them or reassign them with changing needs in the department. We really didn’t want to be stuck with systems that could only run a single assay, so having multiple/ flexible systems was very important to us.
Luigi Covelli: Vendor service and support are very important to us. As far as the instrument itself, purchasers should look for robustness, expandability to accommodate future needs, integration to other systems and operations, and programmability. And, of course, cost.
Hugo Oliveira: In my opinion, the main factors to consider before buying an ALH instrument are its robustness, including materials of construction, and its versatility for accommodating different analytical approaches. Software should also be “global,” meaning it should control more than just the liquid handling system; [it] should perform data processing as well.
Hardik Shah: I have never been involved in the purchase of an ALH system, but I suppose throughput and reliability would be assets.
Alex Nisthal: When we purchased the EVO we were looking for a system that was extendable if our needs changed, and a brand that was highly reputable for precision. We also considered cost and ease of use.
Q: What can automation vendors do to improve their instrumentation, software, interface, or the general user experience?
A: Alex Michel: Personally I’d like to see more and better third-party support from the vendors. We typically go through the vendor of the system for additions to our platforms, but sometimes they don’t have the exact tools that we’re looking for and it can be difficult to implement a third-party solution. Having more options and easier solutions would give us more flexibility with how we use our automation.
Luigi Covelli: Increasing the pipette capacity volume has been an issue for the majority of the ALH instruments on the market. Only a few vendors have excelled on this factor, as well as on liquid level detection. Software and interface in general [have] a steep learning curve regardless of new software upgrades and advancements. Providing additional support for the life of the ALH would increase interest and help overcome the initial “scare” of users stepping into automated robotics.
Hugo Oliveira: In the case of flow analysis, I think the instrumentation is quite well-established and some improvements can be made through innovative designs and the use of state-of-the-art materials in the different components of the manifold. An example of this might be polymeric materials with enhanced chemical resistance. Another key point is software, which is particularly important for the applications involving wet chemistry assays. In this case, it is clear that current data processing and calculation for the generated results in commercially available software are not in the same stage of development as for other instrumentation that works with similar analytical outputs, such as chromatographic methods.
Hardik Shah: Liquid sensing errors are very common when working with plasma samples. This requires constant observation to minimize that error. Sometimes we see “pipette recognition errors” as well, find bubbles in the liquid delivery lines, or pipette tips are off from their intended targets. Improvements in those areas would be most welcome. Another area where improvement is needed is in software and method development. Methods are way too complicated and difficult for average users.
Alex Nisthal: Improvements in reliability and precision are always welcome. As for software and the general user experience, having two versions of the accompanying software could perhaps be useful. The advanced version would be the current version with many, many options to tinker with. The basic version might be useful for novice users who simply want to re-array liquids or something similarly easy. I believe Tecan has something like this now, but I haven’t tried it. Second, improvements to separating the labware and robot vectors of different users and labs are needed.