sepal: identifying transcript profiles with spatial patterns by diffusion-based modeling.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Spatial omics technologies generate large datasets, but analyzing spatial signals presents challenges.
• Distinguishing true spatial patterns from background noise in transcriptomics data is crucial.
• Existing methods often rely on statistical hypothesis testing, prompting the need for alternative strategies.

Purpose of the Study:

• To develop a novel computational approach for analyzing spatial transcriptomics data.
• To identify genes exhibiting significant spatial expression patterns.
• To offer a computationally efficient alternative to existing methods.

Main Methods:

• Simulating the diffusion of individual transcripts to identify genes with spatial patterns.
• Developing an open-source Python package with a command-line interface (CLI).

Main Results:

• The proposed method successfully identified genes with distinct spatial profiles in real data.
• Identified genes are involved in key biological processes and are characteristic of specific cell types.
• The method demonstrated better time performance on multi-core systems and was less influenced by gene expression levels compared to existing approaches.

Conclusions:

• The novel transcript diffusion simulation method effectively extracts spatial patterns from transcriptomics data.
• This approach provides a valuable tool for analyzing complex spatial omics datasets.
• The open-source implementation facilitates broader adoption and application in biological research.