Automated liquid handlers have enabled dramatic improvements in the efficiency of screening the efficacy of potential new drug candidates, but just like a map they need to be calibrated against a universal ‘compass’. Molecules are often selected for further development from high throughput screening (HTS) campaigns after just a single interrogation of a complex biological system. Many factors influence the success of these screening campaigns, and the reliability of the assays to correctly identify promising drug candidates is paramount if pharmaceutical companies are to avoid missing new drug candidates or avoid taking false positives into the next phase of testing. Determining the variability in results stemming from the assay itself requires assessing a combination of liquid handling, biological and random variability sources. The liquid handling component of the overall variability is often underappreciated or ignored, yet can have a significant impact on the outcome of an assay. Many laboratories rely on precision alone to estimate the quality of the liquid handling. However this study illustrates how optimization of both liquid handling accuracy and precision is critical to assay performance. The study was designed to examine: (1) liquid handling performance and (2) the effect of liquid handling variability on in vitro biochemical assays. Specifically, protein binding (streptavidin) and enzyme (alpha-galactosidase) assays were investigated by making small and deliberate changes to delivered volumes of critical assay components. The effect of the liquid handling variability was then measured via inhibitor potencies and assay performance characteristics such as Z-factor, signal-to-background and variability.
The Artel MVS® Multichannel Verification System was chosen as the tool for evaluating liquid handling performance because it is easily adaptable to automated liquid handling systems and it effectively determines both precision and accuracy. Furthermore, Artel Verification plates were compared to five alternate off-the-shelf plate types used for liquid handling QC. Dramatically different accuracy and precision profiles were obtained which could lead to improper calibration of a liquid handler.
Finally, this study demonstrates how an improperly calibrated liquid handler (LH) can affect the performance of model biological assays. Even small changes in dispense volumes can affect inhibitor potency (IC50) values. Ultimately, proper LH calibration is necessary for efficient assay transfer points, such as from assay development to automation validation, primary screening to confirmation screening, lead generation to lead optimization, or simply replacing one liquid handler with another if failure occurs during a campaign.