Spatial motifs reveal patterns in cellular architecture of complex tissues

Area of Science:

• Genomics and Computational Biology
• Neuroscience
• Systems Biology

Background:

• Cellular spatial organization is vital for tissue function and disease, like cancer.
• Spatial transcriptomics offers gene expression and spatial context but current analysis methods overlook higher-order cell type associations.
• Understanding complex tissue development and function requires analyzing these higher-order associations.

Approach:

• Developed a novel algorithm for detecting spatial motifs in neighborhood graphs.
• Utilized uniform sampling of paths from graphs and a motif-finding algorithm inspired by DNA sequence analysis.
• Validated the method using synthetic data for accuracy and sensitivity.

Key Points:

• The method identifies spatial arrangements of cell types that occur more frequently than expected by chance.
• Applied to mouse retinal and hypothalamic data, it uncovered previously unrecognized cell arrangement patterns.
• Identified spatial patterns where cells exhibit distinct gene expression, suggesting functional significance.

Conclusions:

• The novel method effectively detects spatial motifs, illuminating complex tissue organization.
• Reveals novel insights into neural tissue complexity and generates testable hypotheses.
• Enhances understanding of cell-type interactions and their functional roles in biological systems.