High-confidence structural predictions of extrachromosomal DNA with ecDNAInspector

Area of Science:

• Genomics
• Cancer Biology
• Bioinformatics

Background:

• Extrachromosomal DNA (ecDNA) are circular genomic elements outside chromosomes that amplify oncogenes and promote tumor progression.
• Despite their role in aggressive disease, ecDNA formation, propagation, and structure remain poorly understood.
• Existing computational tools struggle with accurate identification of large or complex ecDNA structures.

Purpose of the Study:

• To introduce ecDNAInspector, a novel computational framework for systematically assessing the confidence of ecDNA predictions.
• To improve the interpretability of ecDNA identification and facilitate associations with clinical features.
• To provide a scalable, data-driven approach for characterizing ecDNA structure.

Main Methods:

• Development of ecDNAInspector, a computational framework leveraging short-read whole genome sequencing (WGS) data.
• Systematic assessment of confidence in ecDNA predictions from existing inference tools.
• Application to population-scale cohorts, including 250 breast cancer samples, with orthogonal omic data and experimental validation.

Main Results:

• ecDNAInspector accurately identifies high-confidence ecDNA calls, enhancing interpretability.
• Demonstrated associations between ecDNA structure and molecular subgroups in breast cancer.
• Validated findings using orthogonal omic data and experimental characterization.

Conclusions:

• ecDNAInspector provides a robust method for characterizing ecDNA structure from WGS data.
• This framework enables integrative studies on the clinical and biological impact of ecDNA.
• Facilitates understanding of non-Mendelian oncogene amplification and inheritance in cancer.