Machine learning-optimized targeted detection of alternative splicing.

Area of Science:

• Molecular Biology
• Genomics
• Bioinformatics

Background:

• RNA sequencing (RNA-seq) is a standard tool for transcriptome analysis.
• However, RNA-seq has inherent biases that limit the accurate detection and quantification of alternative splicing events.

Purpose of the Study:

• To develop an efficient targeted RNA-sequencing method for enhanced detection and quantification of alternative splicing.
• To address the limitations of standard RNA-seq in capturing splicing variations.

Main Methods:

• Introduced Local Splicing Variation sequencing (LSV-seq), a targeted RNA-seq approach.
• LSV-seq uses multiplexed reverse transcription with primers designed by the Optimal Prime machine learning algorithm.
• Leveraged deep learning predictions to target splicing events with low coverage in existing RNA-seq data.

Main Results:

• LSV-seq demonstrated high on-target capture rates and concordance with standard RNA-seq.
• The method requires significantly lower sequencing depth compared to conventional RNA-seq.
• Successfully identified hundreds of novel tissue-specific splicing events by targeting low-coverage events in GTEx data.

Conclusions:

• LSV-seq provides a highly sensitive and efficient method for quantifying alternative splicing events.
• This technique enables high-throughput analysis and discovery of splicing variations.
• LSV-seq overcomes limitations of standard RNA-seq for comprehensive splicing analysis.