Analysis of Multimodal Data Integrating Genomics and Spatial Information

Abstract

The advancement of modern computational technologies has paved the way for the processing of high-dimensional data. Benefiting from this, current research in the bioinformatics field has been utilizing sophisticated deep learning architecture to understand the complex human biological system. Spatial Transcriptomics (ST), a popular recent technology, generates multimodal data by integrating imaging data as spatial context, thereby providing extra spatial information to the high-dimensional gene expression profiles. This multimodal, high-dimensional data demands analysis from methods that jointly analyze both the sequencing and imaging data. One critical task that those methods are trying to improve is Spatial Domain identification (SDI). If SDI is well executed, the identified spatial domain can offer valuable biological insights, ultimately facilitating improved diagnosis and treatment strategies for diseases. This project will delve into spatial transcriptomics and its associated analyses to improve our capability of interpreting such data.