Inter-cellular interactions are known to play an important role in response to cancer therapies. To study interactions between cells, the spatial organization of cells in tumor tissue is profiled using spatial transcriptomic technologies that enable visualization and quantification of mRNA transcripts at the single-molecule level. In order to derive biological insights from these images, one crucial step is detecting the membranes (boundaries) of cells. In this project, we will leverage machine learning and deep learning methods developed for segmenting images and detecting objects in computer vision. We will adapt these techniques for the goal of segmenting cells in tissue imaging using assumptions such as preserving geometrical properties or encoding relationships between neighboring pixels.
Lab: Computational Cancer Biology Laboratory
Direct Supervisor: Elham Azizi
Qualifications: Machine learning or Deep learning; Programming in Python