Generative Design of Cell Type-Specific RNA Splicing Elements for Programmable Gene Regulation.

Area of Science:

• Molecular Biology
• Genetics
• Bioinformatics

Background:

• Controlling gene expression in specific cell types is crucial for biological research and developing new therapies.
• Existing methods for cell-specific gene regulation face challenges in scalability and applicability across different cell types.

Purpose of the Study:

• To introduce SPICE (Splicing Proportions In Cell types), a novel framework for programmable, cell type-specific gene regulation.
• To utilize alternative RNA splicing as a modality for precise control of gene expression.
• To demonstrate the potential for engineering cell type-specific splicing for research and therapeutic applications.

Main Methods:

• Developed a massively parallel reporter assay (MPRA) with 46,372 human sequences.
• Profiled exon skipping patterns across 43 cell lines from 10 distinct lineages.
• Trained deep learning models to predict splicing and generate synthetic sequences with programmed splicing patterns.

Main Results:

• Discovered widespread cell type-specific alternative RNA splicing events.
• Created predictive models capable of forecasting splicing in new cellular contexts.
• Generated synthetic sequences exhibiting programmed, cell type-specific splicing behaviors.
• Engineered sequences that selectively splice in cells with specific oncogenic mutations.

Conclusions:

• SPICE offers a scalable and generalizable strategy for dissecting alternative splicing regulation.
• This framework enables the engineering of alternative splicing for precise gene expression control.
• SPICE has significant potential for advancing both basic biological research and therapeutic interventions.