Abstract
BACKGROUND: Advances in bioinformatics have revealed the potential of whole exome sequencing (WES) for copy number variations (CNVs) detection. This study aimed to evaluate whether WES can replace low pass copy number variation sequencing (CNV-seq) for the detection of CNVs in clinical prenatal diagnosis. METHODS: Based on a cohort of 1055 fetuses tested concurrently with trio-WES and CNV-Seq, we performed a systematic comparison of both methods for the detection of aneuploidies, mosaicism, and pathogenic/likely pathogenic (P/LP) CNVs. RESULTS: Both WES and CNV-Seq concordantly identified 11 aneuploidies, 4 mosaicism, and 29 P/LP CNVs. The aneuploidies comprised eight cases of trisomy 21, two cases of 45,X, and one case of 47,XXX. Four mosaicisms consisted of two whole-chromosome mosaicism (WCM) and two segmental aneuploid mosaicism. The two platforms demonstrated comparable accuracy in copy number (CN) detection for both aneuploidies and whole-chromosome mosaicism. However, WES showed comparatively lower precision than CNV-Seq in detecting CN alterations within segmental aneuploid mosaicism. For P/LP CNVs, both platforms demonstrated full concordance in the quantity and pathogenicity classification, although minor discrepancies were observed in their genomic locations and sizes determination. The Jaccard similarity coefficient values ranged from 57.9% to 99.7% (mean±SD, 90.2%±9.5%). Notably, only three cases fell below the 80% threshold, 1q21·1q21.2 deletion (66.0%), 1q21·1q21.2 duplication (57.9%) and 2q32·2q32.1 deletion (78.2%). After excluding these three samples, the similarity improved to 81.3%-99.7% (mean± SD, 92.7%±5.4%). CONCLUSION: WES shows comparable detection of aneuploidies, WCM, and P/LP CNVs to CNV-Seq, supporting its use as a first-tier prenatal diagnostic approach. However, caution remains for segmental mosaicism and certain regions with reduced precision. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12967-026-07899-4.