SVCFit: Inferring structural variant cellular fraction in tumors.

Area of Science:

• Genomics
• Cancer Biology
• Bioinformatics

Background:

• Tumor evolution is driven by cellular mutation and selection, leading to subclones competing for dominance.
• Structural variants (SVs) offer insights into tumor evolution, genomic instability, and late-stage tumor development.
• Accurate identification and quantification of subclones are crucial for understanding tumor heterogeneity.

Purpose of the Study:

• To present SVCFit, a scalable computational method for estimating the cellular prevalence of somatic structural variants (deletions, duplications, inversions).
• To demonstrate improved accuracy in cellular prevalence estimation by incorporating distinct read patterns for different SV types.
• To validate SVCFit's performance using simulated data and patient-derived metastatic samples.

Main Methods:

• Developed SVCFit, a scalable algorithm to estimate cellular prevalence of somatic deletions, duplications, and inversions.
• Incorporated distinct read patterns specific to each structural variant type to enhance estimation accuracy.
• Validated the method on simulated datasets and real patient metastatic samples with known mixture proportions.

Main Results:

• SVCFit significantly improves the accuracy of structural variant cellular prevalence estimation compared to state-of-the-art methods (p<0.05).
• The method achieves high accuracy without requiring prior knowledge of tumor purity, overcoming limitations of inaccurate purity estimates.
• SVCFit demonstrates robust performance across simulated and real-world cancer genomic data.

Conclusions:

• SVCFit provides a more accurate and accessible approach for structural variant-based clonal reconstruction.
• The scalability and speed of SVCFit make it suitable for large cohort analysis using cost-effective bulk whole-genome sequencing (WGS).
• This advancement enhances the study of tumor evolution, genomic instability, and cancer subclonal architecture.