Spatial-omics technologies such as 10x Visium, Xenium, and MERSCOPE are revolutionizing our understanding of tissue organization by mapping gene expression profiles in the spatial context. However, challenges remain particularly in the accuracy of cell segmentation and cell-type annotation, which can be hindered by noisy or incomplete data.
This project will explore cutting-edge artificial intelligence (AI) and machine learning techniques to improve spatial transcriptomics analysis by integrating morphological features from H&E-stained images with spatial gene expression profiles.
By leveraging convolutional neural networks (CNNs), graph neural networks (GNNs), or multimodal learning frameworks, this project aims to improve cell-type annotation by integrating histological context with spatial gene expression, enabling more accurate and biologically meaningful interpretation of spatial-omics data.