Crucial parameters for precise copy number variation detection in formalin-fixed paraffin-embedded solid cancer samples

Copy number variations (CNVs) play a crucial role in cancer diagnostics and prognostics, potentially impacting treatment decisions. Ultra-low-pass whole-genome sequencing (ULP-WGS) has emerged as a promising alternative to array-based methods for CNV detection, especially in formalin-fixed paraffin-embedded (FFPE) samples. However, sequencing biases and sample heterogeneity necessitate the optimization of CNV detection tools for FFPE sample-derived data. This study evaluates three open-source CNV callers (CNVpytor, ichorCNA, and WisecondorX) using ULP-WGS and compares their performance against a single nucleotide polymorphism (SNP) array. Our results demonstrate that under optimal experimental conditions, ichorCNA and WisecondorX achieved equal detection of true positive results, with reduced false positive results compared to the SNP array. The SNP array detection pattern differed somewhat from that of the CNV callers, while ichorCNA and WisecondorX had the most comparable detection pattern. We highlight the importance of (pre-)analytical parameters such as neoplastic cell content, sequencing coverage, and bin size selection on CNV detection accuracy. Our findings support the adoption of ULP-WGS-based CNV detection as a robust alternative to SNP arrays, with WisecondorX emerging as the most suitable tool for clinical implementation.