(917) Non-invasive prenatal detection of copy-number variations based on maternal cfDNA fragmentomics

Objective: Cell-free DNA (cfDNA) based prenatal screening determines the risk of genetic abnormalities using cfDNA of fetal origin circulating in maternal plasma. Current methods detect chromosomal scale aberrations or very large ( >1 Mb) copy number variants (CNVs), missing smaller ones which account for up to 15% of developmental disorders in children. To identify shorter CNVs, we developed a whole genome sequencing (WGS) approach that utilizes both read coverage and fragmentomics features.

Study Design: Our cohort includes 10 women with singleton pregnancies for whom WGS was performed on cfDNA and maternal gDNA. Seven out of the ten pregnancies were sampled in the 1st trimester, with a median fetal fraction (FF) of 7.7%. CNVs causing genetic disorders are typically de novo and dominant, or, in the case of a recessive condition, inherited alongside another pathogenic variant from the other parent, resulting in compound heterozygosity. Therefore we focused on the detection of de novo CNVs >0.1 Mb and smaller CNVs that could cause compound heterozygosity. Paternal CNVs are detected in a similar manner to de novo variants and so were also included in the analysis.

Results: We implemented an algorithm which calculates a Z-score for the difference in cfDNA read lengths across the genome. Two deletions, of size 690 KB and 209 KB were detected in two families. Validation with fetal WGS confirmed the deletions as true positives, with no other relevant deletions found. In a third family, the mother is a known carrier of a pathogenic ~5 KB deletion in the CERS3 gene, and the father carries a pathogenic SNV in the same gene. Our algorithm successfully determined that the fetus did not inherit the maternal CNV (Figure 1).

Conclusion: We present a method to identify CNVs in a cfDNA-based prenatal screening and show that combining read length with coverage data allows reliable CNV detection. Importantly, we demonstrate our method on mostly 1st trimester pregnancies with a low FF, representing a clinically relevant scenario, thus enhancing the potential scope and resolution of noninvasive prenatal screening.