Deciphering splicing heterogeneity at single-cell resolution by SCSES.

Area of Science:

• Computational Biology
• Genomics
• Transcriptomics

Background:

• Alternative splicing (AS) is crucial for cellular diversity.
• Single-cell RNA sequencing (scRNA-seq) is widely used but faces challenges in accurately characterizing AS due to noise and dropout.
• Existing methods struggle to fully capture splicing variations at the single-cell level.

Purpose of the Study:

• To develop a computational framework, SCSES (Single-Cell Splicing Estimation), to enhance AS profiles in scRNA-seq data.
• To improve the accuracy of splicing change characterization at the single-cell level.
• To uncover novel cell subgroups defined by distinct splicing patterns.

Main Methods:

• Developed SCSES, a data diffusion-based computational framework.
• SCSES infers and imputes missing splicing events by leveraging information from similar cells and splicing events.
• Validated SCSES through systematic simulation studies and application to diverse real-world scRNA-seq datasets.

Main Results:

• SCSES demonstrated superior performance over existing algorithms in recovering percent spliced-in (PSI) values and splicing diversity.
• Application of SCSES revealed significant splicing heterogeneity not detectable by conventional gene expression clustering.
• Identified distinct cell subgroups characterized by unique splicing signatures.

Conclusions:

• SCSES is a robust computational tool for accurate single-cell splicing analysis.
• The framework effectively enhances AS profiles, enabling the discovery of novel biological insights.
• SCSES is versatile and applicable across various biological contexts, species, and sequencing technologies.