Spatial Patterning Analysis of Cellular Ensembles (SPACE) discovers complex spatial organization at the cell and...

Area of Science:

• Computational Biology
• Bioinformatics
• Spatial Transcriptomics

Background:

• Tissue organization and spatial patterns of cellular elements are crucial for physiological and pathological processes.
• Discerning complex spatial patterns involving multiple biological elements in tissues is a significant challenge.
• Existing methods often struggle with the complexity of high-plex spatial data.

Purpose of the Study:

• To introduce Spatial Patterning Analysis of Cellular Ensembles (SPACE), an R package designed for the analysis of high-plex spatial data.
• To provide tools for detecting and characterizing diverse spatial patterns, including co-occurrence, context-dependent associations, gradients, and orientations.
• To enable the identification of significant patterns within single images and across different sample groups.

Main Methods:

• SPACE utilizes an information-theoretic framework to explore all possible ensembles of tissue elements (single, pairs, triplets, etc.).
• It ranks ensembles based on the strength of spatial patterning, distinguishing from random assortment or differences between sample groups.
• The package is compatible with various data modalities recording values at fixed spatial coordinates (pixels or voxels).

Main Results:

• SPACE successfully detects and characterizes complex spatial patterns in biological tissues.
• Validation using murine lymph node images with defined ground truth demonstrates SPACE's advantages.
• Application to diverse datasets, including tumors and granulomas, reveals novel spatial organizational patterns.

Conclusions:

• SPACE offers a robust and versatile framework for analyzing high-plex spatial data, uncovering intricate tissue organization.
• The R package facilitates intuitive interpretation of detected spatial patterns, aiding biological discovery.
• SPACE enhances our ability to understand the role of spatial patterning in health and disease across various biological contexts.