Machine learning-based estimation of spatial gene expression pattern during ESC-derived retinal organoid development

Summary Organoids, which can reproduce the complex tissue structures found in embryos, are revolutionizing basic research and regenerative medicine. In order to use organoids for research and medicine, it is necessary to assess the composition and arrangement of cell types within the organoid, i.e., spatial gene expression. However, current methods are invasive and require gene editing and immunostaining. In this study, we developed a non-invasive estimation method of spatial gene expression patterns using machine learning. A deep learning model was trained with an encoder-decoder architecture on a dataset of retinal organoids derived from mouse embryonic stem cells. This method successfully estimated spatially plausible fluorescent patterns with appropriate intensities, enabling the non-invasive, quantitative estimation of spatial gene expression patterns within each tissue. Thus, this method could lead to new avenues for evaluating spatial gene expression patterns across a wide range of biology and medicine fields. Highlights A non-invasive estimation method of spatial gene expression pattern is proposed A CNN architecture is employed to convert a phase-contrast to fluorescence image The method was trained on a dataset of mouse ESC-derived retinal organoids Spatially plausible patterns of Rx gene expressions were successfully estimated