Harnessing the potential of spatial statistics for spatial omics data with pasta

Area of Science:

• Computational Biology
• Spatial Omics
• Statistical Analysis

Background:

• Spatial omics enables molecular profiling of cells within their native tissue environment.
• Imaging-based and high-throughput sequencing-based spatial omics yield distinct data modalities.
• These data types correspond to point patterns and lattice data in spatial statistics.

Purpose of the Study:

• To demonstrate the utility of spatial statistics for analyzing diverse spatial omics data.
• To quantify biological phenomena ranging from local gene expression to tissue organization.
• To provide a framework for addressing scientific questions in breast cancer research using spatial statistics.

Main Methods:

• Application of spatial statistics concepts, including window sampling, homogeneity, and weight matrix construction.
• Exploratory metrics for analyzing cellular co-localization and gene co-expression.
• Utilizing R and Python packages for spatial data analysis, exemplified by the 'pasta' vignette.

Main Results:

• Spatial statistics effectively quantifies biological phenomena from spatial omics data.
• Exploratory metrics can address specific scientific questions in breast cancer, such as cellular and gene co-localization.
• The importance of technical concepts like window sampling and weight matrix construction is highlighted.

Conclusions:

• Spatial statistics provides a versatile framework for understanding spatial omics data.
• The methods discussed are applicable to various biological questions, including those in oncology.
• The 'pasta' vignette serves as a practical resource for implementing spatial statistical analyses.