Systematisk fenotypisk profilering av celler i farmakologisk säkerhetsbedömning
Tidsperiod: 2020-12-01 till 2023-12-31
Projektledare: Ola Spjuth
Medarbetare: Carolina Wählby, Robert Fredriksson, Jordi Carreras Puigvert
Budget: 2 280 000 SEK
Assessing safety of early candidates in drug discovery is currently to a large extent based on animal studies, requiring large amounts of animals every year. The purpose of this project is to validate our previous findings that phenotypic profiling with high-content imaging can be incorporated into the standard in vitro drug safety pipeline, and combined with AI modeling enable identification of additional safety concerns and thereby reduce the number of compounds that unnecessarily proceed to animal testing.We propose a three year project comprising an interdisciplinary team supported by experts in image analysis, AI/Deep Learning, and cell biology with access to a robotized cell profiling laboratory. We will carry out validation experiments of a large set of compounds, producing unique and high-quality data resources. Based on our recent methods using AI/Deep Learning, we will train AI models to predict safety issues with high confidence based on these datasets. We will also pursue studies to integrate data from different labs.The proposed methodology offers a cheap and robust way of improving in vitro safety pharmacology. By developing and publishing cell and imaging protocols, making data and analysis pipelines available as Open Access, and developing blueprints for how to automate lab protocols and data analysis, we ensure that the developments can be easily integrated into pharmaceutical companies and thereby directly contribute to a reduction in animal experiments.