Gene Fusion Detection in Long-Read Transcriptome Datasets from Multiple Cancer Cell Lines.

Area of Science:

• Genomics
• Cancer Research
• Bioinformatics

Background:

• Fusion genes are critical cancer biomarkers due to oncogenic abnormal proteins.
• Long-read RNA sequencing (long-read RNA-seq) enables full-length mRNA transcript sequencing for fusion gene detection.
• High error rates in long-read RNA-seq pose challenges for accurate fusion gene identification.

Purpose of the Study:

• To develop an improved tool for detecting fusion genes in long-read RNA sequencing data.
• To enhance the accuracy of fusion gene detection despite inherent sequencing errors.

Main Methods:

• Incorporated novel steps: breakpoint anchoring to exon boundaries, realignment of unaligned regions, and breakpoint clustering.
• Evaluated tool accuracy by comparing it against established tools like JAFFAL and FusionSeeker.
• Utilized long-read RNA sequencing datasets from cancer cell lines.

Main Results:

• The developed tool demonstrated superior performance in detecting fusion genes compared to JAFFAL and FusionSeeker.
• Identified potential novel fusion genes through consistent detection across multiple tools and datasets.
• Validated the tool's effectiveness on real-world cancer cell line data.

Conclusions:

• The new tool effectively detects fusion genes in long-read RNA sequencing data.
• This advancement aids in identifying crucial cancer biomarkers.
• The tool's performance was confirmed across different cancer cell line datasets.