Statistical methods for spatial Multiplexed Ion-Beam Imaging data analysis

Author(s): Shiheng Huang,Pratheepa Jeganathan,Jamie McNicol

Affiliation(s): McMaster University



This workshop aims to enable users to configure and follow the reproducible research workflow of statistical methods in characterizing the microenvironment of pathological tissues at the molecular level. In this context, an important goal is to explain the heterogeneity observed in disease specimens through investigating the spatial compartments present in the tissue. We will demonstrate the use of latent Dirichlet allocation (LDA) to identify latent cell phenotype communities as well as tessellation to investigate cell phenotype spatial dependence. This workshop is structured around three modules: cell phenotype annotation, topic analysis, and spatial dependence analysis. We will showcase an interactive R Shiny app that can be tailored to individual analyses to annotate cell phenotypes with agglomerative hierarchical clustering and the computational workflow for the LDA and local spatial dependence measures to provide insights alongside the clinical metadata.