NoVaTeST: identifying genes with location-dependent noise variance in spatial transcriptomics data.

Area of Science:

• Genomics
• Computational Biology
• Bioinformatics

Background:

• Spatial transcriptomics (ST) enables analysis of gene expression patterns within tissues.
• Current ST analysis tools often assume constant noise variance, potentially overlooking biological signals.
• Variations in noise variance across spatial locations can be biologically significant.

Purpose of the Study:

• To introduce NoVaTeST, a novel framework for detecting genes with location-dependent noise variance in ST data.
• To address the limitations of existing methods that assume constant noise variance.
• To identify novel biological insights from spatial transcriptomics data by analyzing noise variations.

Main Methods:

• NoVaTeST models gene expression as a function of spatial location, allowing for spatially varying noise.
• It statistically compares a spatially varying noise model against a constant noise model.
• Genes exhibiting significant spatial noise variation are identified as "noisy genes".

Main Results:

• NoVaTeST successfully identifies genes with location-dependent noise variance in ST data.
• In tumor samples, the "noisy genes" identified by NoVaTeST are largely distinct from those identified by traditional methods.
• These "noisy genes" offer unique biological insights into tumor microenvironments.

Conclusions:

• NoVaTeST provides a powerful new approach for analyzing spatial transcriptomics data.
• The framework enhances the discovery of biologically relevant spatial patterns by considering noise variability.
• This method has the potential to advance our understanding of complex tissue functions and disease mechanisms.