Abstract

Spatial omics technology offers unprecedented opportunity to study in-situ gene expression and cell-cell interaction. However, spatial omics data also suffer from considerable heterogeneity introduced during data generation processes. In this talk, I am going to present two statistical methods that we have developed to address this issue. The first method called SpaNorm is designed to reduce library size effects in single-sample and multi-slice spatial transcriptomics data. The second method, bgnorm removes background fluorescence intensity in imaging-based spatial proteomics data. The statistical models underlying each method will be discussed, as well as some benchmarking results and potential future works.