Prediction of alternative isoforms from exon expression levels in RNA-Seq experiments

Area of Science:

• Molecular Biology
• Genomics
• Bioinformatics

Background:

• Alternative splicing, polyadenylation, and differential promoter usage generate diverse transcript isoforms with specific biological functions.
• The full range of mammalian alternative transcripts and their regulatory mechanisms remain incompletely understood.
• Advances in second-generation sequencing (RNA-Sequencing) offer powerful new approaches for transcriptome-wide analysis.

Purpose of the Study:

• To develop and validate methods for predicting and quantifying alternative RNA isoforms from RNA-Sequencing data based on exon expression levels.
• To enable the analysis of alternative splicing events within and between experimental conditions using existing gene annotations.
• To provide a comprehensive computational framework for analyzing alternative transcript isoforms.

Main Methods:

• Development of an explicit statistical model for predicting alternative isoforms solely from exon expression levels in RNA-Sequencing data.
• Implementation of methods for the relative quantification of known transcript structures.
• Validation of predicted isoforms using RT-PCR and comparison with junction read data and exon arrays.

Main Results:

• Successful prediction and validation of a significant fraction of alternative isoforms from human RNA-Sequencing data.
• Demonstrated correlation between predicted isoforms and junction read information.
• Showcased improved performance of RNA-Sequencing over microarrays for predicting skipped exons.

Conclusions:

• The presented methods offer a comprehensive approach to analyzing alternative splicing and transcript isoform diversity.
• The developed computational tools, available as the open-source R-package Solas, facilitate the study of transcript regulation.
• This work advances the understanding of the mammalian transcriptome and gene expression complexity.