Development of computational methodologies for subcellular omics observations

In recent years, the resolution of spatial omics observations has been improving significantly, and it is becoming possible to obtain comprehensive spatial distributions of molecular profiles inside cells. On the other hand, information analysis technologies that can handle omics profiles with subcellular resolution have hardly emerged yet. We are developing techniques for reconstructing three-dimensional spatial patterns for such observation data by utilizing deep learning techniques. This will enable us to elucidate the molecular profiles of compartments inside the cell caused by liquid-liquid phase separation, etc.