What happens behind the scenes at a leading genomics service provider? An Automation Scientist* with the R&D group at a CRO recently shared some of his insights and experiences with us.
We’re a genomics service company that is committed to furthering our clients’ efforts to uncover genetic elements important to human health and the diagnosis and treatment of disease. We serve the academic, biotech, and pharmaceutical industries, providing expression analysis and genotyping tests, as well as all the newest techniques such as next-generation sequencing (NGS).
More generally, we’re about reliable data, high-quality data—many of our clients keep coming back because they know that they will get the data they need to move forward with their projects in a timely fashion.
I’m an Automation Scientist with the R&D group. The R&D group works with clients who want to create custom assays and I automate those assays, writing software and running the automated liquid handlers. I also work with the internal production group when they want to automate a manual assay.
One way is through investment in automation, which is what I take care of. We have five automated liquid handlers and our goal is to get as much of each workflow automated as we can. Our most popular assays are pretty much fully automated, which helps tremendously with consistency and reproducibility.
Another way is through our quality systems—our services are conducted in a CLIA-certified lab which supports GLP compliance, and our quality systems follow CLSI (Clinical and Laboratory Standards Institute) guidelines.
My part in ensuring quality—which takes up the majority of my time—involves the IQs, PQs, and OQs of our automated systems.
IQs are the instrument qualifications, which is what the service engineer does when he/she installs the instrument and verifies that the instrument is operating up to specifications upon installation. When new instruments come in, I review the IQs.
PQs are the performance qualifications and can take up to 75% of the time needed to get an assay from R&D into production. It involves validating the assay to ensure that it is able to achieve the goal within the appropriate specifications. Because we follow CLIA guidelines, we typically develop a validation plan that includes running samples that are representative of the test articles. For example, we don’t want to run validation assays with perfect, high-quality samples if we know that the client will be providing us with FFPE (formalin-fixed, paraffin-embedded) samples. Instead we do our validations using degraded DNA and make sure our assays perform under these more real-world conditions.
We also try to include variability in our validation assays to mimic lot-to-lot differences in reagents and consumables as well as operator-to-operator variability, and we observe general day-to-day assay consistency. We look at all these metrics to ensure that our assays are highly robust, and so that we can fully characterize assay accuracy and repeatability, the upper and lower limits of our assay’s detection, and the assay’s dynamic range—all of your typical assay performance metrics.
OQs are the operational qualifications that demonstrate that the automated liquid handlers are performing within acceptable specifications—basically, to prove that when we think we’re pipetting one microliter, we are actually pipetting one microliter. At my company, we perform OQs once a year as we find that our liquid handlers typically stay within specifications within this time frame.
Recently, we’ve turned to Artel’s Services to perform our yearly OQs. We use an outside vendor to be able to show any quality or compliance auditors that a third party has independently calibrated and qualified our liquid handlers.
One reason we’ve chosen Artel is because, at the price point, we get a lot of data and value. Our previous calibration service only offered calibration at two different volumes—5 µl and 200 µl—but Artel provides us with actual calibrations over dynamic pipetting ranges and under various circumstances, such as tip types and sample types. To ensure that I’ll be getting the testing I need I just pick up the phone and talk to Dana Campbell, our Artel Laboratory Applications Specialist. We review the different assay conditions, instruments, and labware we use and Dana develops a custom testing plan that is optimized for our needs.
Obviously, with this rich data set I feel much more confident in my instruments, my ability to deliver consistent, quality data to our clients, and my ability to defend our systems to any auditors. But I can also use the data to optimize the performance of my liquid handlers. For example, I can tell that for dispensing between 1 – 10 µl, I’ll get the best performance with specific offset and calibration values, but that for volumes greater than 10 µl I should use a different offset and calibration values. With this approach, I can often get better performance than the manufacturer’s specifications, allowing us to really get great pipetting consistency and accuracy.
Another reason I like Artel is because Dana and the Artel Liquid Handling Service team are a great general resource. We were recently considering developing a new assay using a viscous, oily solution—almost all of our assays are aqueous—and I was able to talk to Dana about the best ways to calibrate our liquid handlers for this new-to-us condition. Even though we ended up not using those conditions, it was great to have Dana’s feedback on how to handle the situation and to know that it was something Artel already had experience with.
I keep hearing from customers that they come back to us because of the quality of our data and analysis. Our clients know they can depend on our data—it’s good, reliable, and on time—which I like to think of as a vote of confidence in my work. Through the automation and quality systems I maintain, with help from Artel’s Liquid Handling Services, I help our clients quickly get the answers they need to make critical decisions based on genomics studies.
*Kept anonymous to preserve the confidentiality of our customer.