Benchmarking copy number variation detection with low-coverage whole-genome sequencing.

Abstract

Low-coverage whole-genome sequencing (lcWGS) provides a cost-effective method for genome-wide copy number variation (CNV) profiling, yet its technical limitations and analytical variability require systematic evaluation. We benchmarked five CNV detection tools using simulated and real-world datasets, focusing on sequencing depth, formalin-fixed paraffin-embedded (FFPE) artifacts, tumor purity, multi-center reproducibility, and signature-level stability. Our results demonstrate that ichorCNA outperformed other tools in precision and runtime at high purity (≥50%), making it the optimal choice for lcWGS-based workflows. Prolonged FFPE fixation induced artifactual short-segment CNVs due to formalin-driven DNA fragmentation, a bias none of the tools could computationally correct, necessitating strict fixation time control or prioritization of fresh-frozen samples. Multi-center analysis revealed high reproducibility for the same tool across sequencing facilities, but comparisons between different tools showed low concordance. Copy number features extracted by the Wang et al. method exhibited superior stability across conditions compared with the Steele et al. method and the Tao et al. method. This study establishes actionable guidelines for lcWGS: prioritize ichorCNA (ensuring ≥50% tumor purity), optimize FFPE protocol, and use Wang et al. features to ensure robust copy number profiling in precision oncology.