iIMPACT

iIMPACT is a novel method which integrates image and molecular profiles for spatial transcriptomics analysis. It is a multi-stage statistical method for spatial domain identification and domain-specific spatially variable gene detection. It utilizes an interpretable Bayesian finite mixture model for analyzing the cellular spatial organization and a regression model for domain-specific spatially variable gene analysis. Compared with existing methods for spatial transcriptomics data analysis, iIMPACT fully leverages the biological features from histology image, which causes compromised algorithm accuracy and the interpretability of histological characterization of single cells on domain level. iIMPACT provides a highly accurate and interpretable clustering approach to reveal cellular spatial organization and functional gene landscape from spatial transcriptomics data.

Contents:

• Installation
• Example - Human Breast Cancer 10x Visium Data
  • Load Data
  • Process Data
  • Spatial Domain Identification
  • Domain-specific Spatially Variable Gene Detection
• Example - Mouse Visual Cortex STARmap Data
  • Load Data
  • Process Data
  • Spatial Domain Identification
• Environment