SCISSOR: a framework for identifying structural changes in RNA transcripts.

Area of Science:

• Genomics and Bioinformatics
• Computational Biology
• Molecular Biology

Background:

• High-throughput sequencing, particularly RNA-sequencing (RNA-seq), enables detailed analysis of RNA transcripts.
• Existing computational tools for identifying mRNA variations like alternative splicing face analytical challenges.
• There is a need for robust methods to discover aberrant RNA transcript structures.

Purpose of the Study:

• To propose a novel computational framework for unbiased and robust discovery of aberrant RNA transcript structures.
• To leverage RNA-seq coverage profile shape changes for identifying structural alterations.
• To enable unsupervised screening of RNA structural abnormalities across cohorts.

Main Methods:

• Developed SCISSOR (Shape changes in selecting sample outliers in RNA-seq), a series of procedures for transforming and normalizing base-level RNA sequencing coverage data.
• The normalization is performed in a transcript-independent manner.
• Implemented a statistical framework for analyzing the transformed data, enabling unsupervised screening of structural alterations.

Main Results:

• SCISSOR successfully recaptures known variants, including splice site mutations in tumor suppressor genes.
• The method identifies novel variants, such as recurrent alternate transcription start sites and complex deletions in 3' UTRs.
• Demonstrates the ability to detect abnormalities that are difficult to identify with existing methods.

Conclusions:

• SCISSOR provides a powerful, unsupervised approach for detecting diverse RNA structural abnormalities from RNA-seq data.
• The framework's transcript-independent nature and focus on coverage profile shape changes offer a robust analytical strategy.
• SCISSOR enhances the discovery of both known and previously unrecognized RNA variants, advancing transcriptomic analysis.