Breaking Barriers in Phenotypic Screens: Full Well Submicron Imaging Meets High Throughput Speeds
Monday, March 11th, 3:00 PM – 4:00 PM, Room 155F
By Matt Boisvert, PhD.
Phenotypic screening through high content imaging (HCI) can give unparalleled insight into cellular function and toxicity, and power drug discovery at scale. However, HCI at speed requires compromise: sacrificing resolution, sample coverage or both. This has limited high content screening’s utility as a high throughput tool, increasing the time between testing and discovery or decreasing the overall quality of the screen. This talk discusses how to get to true high throughput high content screening, with end-to-end workflows with Araceli Endeavor®, an instrument designed to image 96-, 384-, and 1536- well high content plates in under 10 minutes/plate without compromising resolution or sample coverage. Acquired data seamlessly transfers to Araceli Clairvoyance™, an analysis software designed with a combination of AI, machine vision and traditional intensity-based techniques. This throughput is demonstrated with 48+ plates imaged and analyzed in a single workday with full well coverage, capturing ~85% of the well area at submicron resolution, with 0.27µm/pixel. We use cytokine induction and autophagy assays to highlight how full well datasets can eliminate variability, with limited sample coverage generating possibly misleading data. Maximizing data input is essential to train and refine AI algorithms, and looking at the full well ensures adequate coverage of rare events such as particular mitotic timepoints or apoptotic cells. Using the example of double strand break quantification in a genotoxicity assay, we demonstrate the necessity of its submicron resolution in ensuring assay success. This talk further highlights how imaging at this resolution and speed expands the information from an assay beyond primary screening outcomes without added cost, such as leveraging transmitted light for phenotyping and cytotoxicity data (adding 1-3 minutes/plate) and micronuclear detection through common DNA stains. Overall, we discuss the implementation and advantages of an end-to-end solution for phenotypic screening enabling faster drug discovery while maximizing data quality.