Harnessing and Validating Deep Learning-based Image Analysis for Cell Painting as an Unbiased Approach to Phenotypic Drug Discovery

The field of target and drug discovery is rapidly embracing high-content (HC) image-based methods, such as Cell Painting. To address the complexity of tools like CellProfiler, we’ve introduced IKOSA AI, a user-friendly deep learning-based computer vision tool. IKOSA AI eliminates the need for coding knowledge, making it accessible to a broader audience. Validated against JUMP-CP data using StratoMineR, our results demonstrate the model’s reproducibility. We’re also investigating the link between cytoplasm features and phenotypic diversity in Cell Painting. The IKOSA AI and StratoMineR combination offers an automated, coding-free solution, accelerating the generation of biological insights from high-content data and reshaping target and drug discovery. Join us in this transformative journey.