Identifying cell state-associated alternative splicing events and their coregulation.

Area of Science:

• Molecular Biology
• Genomics
• Bioinformatics

Background:

• Alternative splicing significantly impacts cellular identity by diversifying the transcriptome.
• Single-cell RNA sequencing (scRNA-seq) has faced challenges in accurately analyzing alternative splicing due to mRNA recovery limitations.
• Previous studies indicated that low mRNA recovery in scRNA-seq can lead to misinterpretations of cell-to-cell splicing variability.

Purpose of the Study:

• To develop a robust computational method for identifying cell type-specific alternative splicing patterns from scRNA-seq data.
• To overcome the limitations of low mRNA recovery inherent in scRNA-seq.
• To link alternative splicing patterns to cell identity without prior cell clustering or trajectory inference.

Main Methods:

• Introduction of Psix, a novel probabilistic model designed to accurately detect alternative splicing events across single cells.
• Utilizing an autocorrelation-inspired approach to identify splicing patterns correlated with cell identity markers.
• Application of Psix to scRNA-seq data from mouse brain development to analyze exon splicing patterns.

Main Results:

• Psix successfully identified modules of coregulated exons based on their alternative splicing patterns during mouse brain development.
• These identified exon modules were enriched for binding sites of specific neuronal splicing factors.
• Changes in exon splicing correlated with the expression levels of these regulatory splicing factors, revealing regulatory network dynamics.

Conclusions:

• Psix enables confident identification of cell type-specific alternative splicing, offering insights into post-transcriptional regulation.
• The method reveals the regulatory networks controlling cell type-dependent splicing patterns.
• Psix enhances scRNA-seq capabilities, allowing for deeper understanding of cell identity determination through post-transcriptional mechanisms.