Severus detects somatic structural variation and complex rearrangements in cancer genomes using long-read sequencing

Area of Science:

• Genomics
• Bioinformatics
• Cancer Research

Background:

• Long-read sequencing offers advantages for detecting somatic structural variations (SVs) in cancer genomes, including improved mappability and variant phasing.
• Existing long-read SV detection tools struggle with the complex rearrangements and heterogeneity typical of tumor genomes.
• Accurate SV detection is crucial for understanding cancer development and for clinical applications.

Purpose of the Study:

• To develop and evaluate Severus, a novel breakpoint graph-based algorithm for somatic SV calling in long-read cancer sequencing data.
• To assess Severus's performance against existing long-read and short-read SV detection methods.
• To demonstrate Severus's utility in identifying clinically relevant SVs in pediatric leukemia and lymphoma cases.

Main Methods:

• Developed Severus, a breakpoint graph-based algorithm for somatic SV calling using long-read sequencing data.
• Utilized a comprehensive multitechnology cell line panel for method benchmarking.
• Applied Severus to clinical samples of pediatric leukemia/lymphoma.

Main Results:

• Severus consistently outperformed other long-read and short-read SV detection methods in terms of F1 score on a cell line panel.
• Demonstrated that short-read sequencing systematically misses certain SV classes, such as insertions and clustered rearrangements.
• Identified clinically relevant cryptic rearrangements in pediatric leukemia/lymphoma cases that were missed by standard genomic panels.

Conclusions:

• Severus is a robust algorithm for detecting somatic structural variations in complex cancer genomes using long-read sequencing.
• Severus offers superior performance compared to existing methods, particularly for challenging SV types.
• The algorithm has the potential to improve the clinical diagnosis of cancers by revealing previously undetected genomic alterations.