InSituCor: exploring spatially correlated genes conditional on the cell type landscape

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Spatial transcriptomics data offers insights into biological phenomena through gene expression patterns.
• Spatially correlated genes are key to understanding cell-cell interactions and latent biological variables.
• Existing methods often struggle to distinguish true biological correlations from those driven by cell type composition.

Purpose of the Study:

• To introduce InSituCor, a novel toolkit for discovering modules of spatially correlated genes.
• To differentiate true biological spatial correlations from those explained by cell type distribution.
• To facilitate both unbiased and knowledge-driven exploration of spatial gene expression patterns.

Main Methods:

• Development of the InSituCor toolkit for analyzing spatial transcriptomics data.
• Implementation of algorithms to identify gene correlations not explained by cell type landscape.
• Support for unbiased, whole-dataset correlation discovery and targeted gene exploration.

Main Results:

• InSituCor effectively identifies spatially correlated gene modules.
• The toolkit successfully filters out correlations attributable to cell type arrangements.
• InSituCor enables focused analysis, including the evaluation of spatial co-regulation for ligand-receptor pairs.

Conclusions:

• InSituCor provides a powerful approach to uncover meaningful biological signals in spatial transcriptomics.
• By removing confounding factors, InSituCor enhances the efficiency of biological discovery.
• The toolkit advances the analysis of spatial gene expression and cell-cell communication.