Abstract
Sex discordance between cell-free DNA (cfDNA) testing and ultrasound examination is rare but can cause significant patient discomfort and uncertainty. Here, we present two clinical cases where a closer examination of raw sequencing data allowed us to anticipate possible discrepancies caused by the insertion of Y-chromosome regions into the maternal genome. We used Illumina's VeriSeq NIPT Solution v2 and a proprietary bioinformatics pipeline to analyze cfDNA in the maternal bloodstream. Paired-end sequencing data were aligned to the reference genome (hg19). Non-duplicated aligned reads were aggregated into 100-kb bins, adjusted for CG bias, and further aggregated into 5-Mb windows. Z-scores were calculated for autosomes, sex chromosomes, and 5-Mb bins. The two clinical cases were classified as low-risk male fetuses according to the primary statistics (case A: NCV(x) = 0.3; NCV(y) = 40.6; native fetal fraction (FF(i)) = 5.1%, and case B: NCV(x) = -0.3, NCV(y) = 40.7, FF(i) = 10.8%); however, the Y-chromosome-based FF (FF(y)) was significantly lower than the default FF estimate (FF(y) ≅ 2% in both cases). Plots of X and Y chromosome Z-scores for each 5-Mb bin, according to genomic position, identified bins with Z-scores significantly higher than those expected for any pregnancy with a male fetus. The genomic coordinates of these bins overlapped with the amelogenin (AMELY) and protein kinase Y-linked (PRKY) genes, respectively. Amplification of these regions in the DNA isolated from the white blood cells fraction confirmed the presence of Y-chromosome insertions in the maternal genome. This study highlights a new source of discrepancy in cfDNA testing due to maternal genomic variations. These findings suggest the need for improvements to current bioinformatics pipelines to identify and exclude possible maternal perturbations from the classification algorithms used for aneuploidy and sex calls.