Developing Random Field based novel approaches for spatial transcriptomics

点击翻译按钮获取中文摘要

The overall goal of the project is to develop Random Field based approaches to spatially analyze and understand tissue architecture and heterogeneity as well as the communication between different cells and stromal components. The knowledge will gain insights into the potential molecular mechanism related to cell evolution, tissue development, disease progression, and drug resistance. The developed tools will significantly improve our understanding of tissue heterogeneity and development. The characterization and use of subclones and spatial architecture for function analysis will provide a completely different approach to study tissue heterogeneity. Graduate and undergraduate students will work under this project and gain experience in leading-edge research. An undergraduate summer research program on "Computational Systems Biology” will be organized. Systematic analysis methods used to analyze spatial transcriptome data are still in their infancy. Current analysis methods of spatial transcriptome data focus on transcriptional profiling. Novel methods are needed to identify genomic variants and integrate genetic and transcriptional variations. To address the challenges, a hybrid model based on Variational Graph AutoEncoder (VGAE) will be developed to characterize the spatial relationship between the spots and subclones. Hidden Markov Models (HMM) will be used to infer the copy number variations (CNVs) from spatial transcriptomic data. The new approach of identifying subclone and CNVs from spatial transcriptomics through VGAE and HMM is coined as CVAM. A Random Field based computational toolset called SPAT (spatial architectural analysis) is proposed to study the architectural heterogeneity with spatial transcriptomics and scRNA-seq data. SPAT can identify single gene-based spatial biomarkers, analyze spatial distribution patterns of signaling networks, and explore cell-cell interaction using spatial transcriptomics and scRNA-seq data. Specifically, a Gromov-Wasserstein distance and Random Field-based approach will be designed to characterize signaling between adjacent cells and stromal cells. Finally the new computational tools and results will be validated through biological experiments. Software prototypes and the variants and gene biomarkers with spatial patterns will be made publicly available to the research community via a project website at https:/ccsm.uth.edu/NSFSPA.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.