Q: Describe your sample prep workflow(s).
A: Gang Xue: The typical samples involved in pharmaceutical analysis include raw materials, intermediates, active pharmaceutical ingredients, excipients, and formulated drug products. Reference standards are used for quantitative analyses. The majority of sample preparation methods target chromatographic potency and purity analyses, such as for developing processes and performing release or stability testing. We use the Mettler Toledo Quantos; FreeSlate Powderium; and tablet extraction systems such as Sotax TPW3, RTS SoliPrep, Polytron, and ASE (accelerated solvent extraction). Hamilton Microlab and CTC autosampler are also used for sample dilution.
Dhara Patel: We employ two types of automation. We use what I classify as full-automation for our compound screening, or HTS. This takes more effort to set up and validate the entire process, but minimal manual intervention is required once the process is validated. We use liquid handlers, plate handlers, automated incubators, and readout instruments like plate readers and automated imagers. The manual intervention required involves stacking plates and placing reagents when an operation starts. The other type of automation we perform is what I classify as semi-automated, which automates only the most labor-intensive and repetitive steps. We use semi-automation for validation assays after HTS and also for any basic science experiment. In this case, to decide whether a process is worth automating, we weigh the effort required to fully validate the steps against performing the steps by hand.
Most of our samples involve cell lines that are treated with cytokine and/or small molecules, staining of cells thereafter, readouts using plate readers, and an automated imager. We are now preparing to automate such processes for primary cells that require growth on air-liquid interface and work with live viruses (BSL2). This requires automation in a biosafety cabinet, which is the main challenge.
Q: What are the principal benefits to your organization of automating sample preparation?
A: Gang Xue: The key benefits include resource savings and enhanced method robustness. Sample preparation and data analysis are the most labor-intensive unit operations in pharmaceutical analyses. An automated system reduces the time for scientists to develop and operate the sample preparation method. A recent trade publication survey reported that sample preparation consumes about 60 percent of a scientist’s time and that approximately 30 percent of out-of-specification (OOS) errors are due to sample preparation.
In a regulated GMP environment, investigations of OOS results generally take significant time and resources. Via the use of automation, we reduced the sample preparation time for a sustained release tablet from 24 hours through manual methods to 15 minutes. An added benefit was the significantly improved precision of the assay. In most cases, automation doesn’t necessarily reduce the overall cycle time of the analysis because of the sequential nature of most automated systems.
Dhara Patel: The biggest advantage of our fully automated process for HTS, especially for a phenotypic cell-based screen, is data quality from consistent timing between plates. This and our data analysis technique enable us to compare data from multiple screens, which for an academic lab like ours that is not a core and can afford to screen smaller libraries at separate times, often months apart, is extremely important. In the long run, being able to hitpick from multiple screens collectively saves a lot of effort invested in hit-validation and further followup. For our semi-automated processes, the biggest advantage or benefit is time savings. When validating any automation process, we always compare to data obtained by performing the experiment by hand. So we can quantify data quality.
Q: What sources of methods or techniques do you rely on most for automated sample preparation?
A: Gang Xue: Usually lab automation-related conferences and trade shows—such as the Pittsburg Conference, Lab Automation Conference (SLAS) in the United States, and Analytica in the EU—are the most valuable sources for innovative automated sample preparation techniques. The precision of the automated methods varies by the scale, throughput, and unit operation. However, the method robustness is mostly reported to be on par with or better than manual methods. The reliability of the instruments can be an issue though. Commercial instruments are typically more reliable than custom-built systems, and smaller systems targeting limited unit operations are better than large, integrated solutions. Additionally, these smaller, targeted systems are more cost-effective and typically more user-friendly, resulting in general lab use, not just use by specialists.
Dhara Patel: For broad-stroke methodology and learning about new methodology, I rely on publications in journals. When starting a new methodology, I look for any algorithms or workflows or examples that may have been developed by the automation equipment company. When troubleshooting, I turn to message boards online. I talk to tech experts from the companies. I also talk with sales representatives, especially to keep up with what’s new in plasticware. When I read journals or magazines, I look for broadstroke ideas. When I read information from vendors, I’m looking for details in methodology. Generally speaking, I think what I read is reliable. However, I always think that the devil is in the details and care needs to be taken when implementing what I read for my own use.
Q: How can vendors of automation equipment and reagents/ disposables improve the quality and consistency of automated sample or standards prep?
A: Gang Xue: Building smaller modular systems that sell to relatively large user bases would help improve the overall robustness of the instruments as well as reduce operational complexity while lowering costs. Providing standardized hardware and software interfaces to allow easy integration (plug and play) would enable both workflow flexibility and reliability. Standardization of consumables would also be welcome. Many of the challenges with robotic operations are related to the different sizes and shapes of the vessels they have to manage.
Dhara Patel: In general, I think that vendors of both automation equipment and reagents and disposables are good at listening to understand what the customer requires. But I think more progress would be made if more attention were given to smaller markets of niche applications. When purchasing equipment, I think what needs to happen more is that the vendor needs to understand what the usage is going to be and discuss equipment options. Many times details of one system (or even a sample preparation platform) are discussed without paying due attention to whether it would be the best choice for the application.