Institute of Information Science, Academia Sinica

The recent development of high-throughput drug profiling (high content screening or HCS) provides a large amount of quantitative multidimensional data. Despite its potentials, it poses challenges for detecting the effectiveness of thousands of drug candidates from a large volume of cell microscopic images. This talk will present graph-based machine learning methods to quantify morphological differences of high-dimensional subcellular features and how to apply the method to discover candidate drugs for genetic disorders from microscopic images of induced stem-cells.