BinDel: detecting clinically relevant fetal genomic microdeletions using low-coverage whole-genome sequencing-based NIPT

AbstractBackgroundClinically pathogenic chromosomal microdeletions (MDs) cause severe fetal genetic disorders such as DiGeorge and Prader-Willi/Angelman syndromes. Motivated by the absence of reliable blood and/or ultrasound screening biomarkers for detecting microdeletion risk during the first-trimester screening, we developed and validated BinDel, a software package to evaluate the risk of clinically pathogenic microdeletions from low-coverage whole-genome-sequencing (WGS)-based NIPT data.ResultsWe used 584 NIPT samples, including 34 clinically pre- and postnatally confirmed microdeletions, to perform a blind evaluation of the BinDel software. In a combined analysis of 34 microdeletion and 50 euploid fetal samples, BinDel correctly identified 25 samples with microdeletions in the ‘blind’ analysis. BinDel had 15 false-positive microdeletion calls, whereas the majority of them were concentrated in a few challenging regions, like NF1 microdeletion region. As a comparison, WisecondorX identified 16 correct microdeletion calls with no false-positive calls. After improving BinDel, 30 microdeletion samples were correctly determined, with a total of three false-positive microdeletion calls. Using simulated fetal microdeletions, we investigated the impact of fetal DNA fraction (FF) and microdeletion region length on BinDel’s microdeletion risk detection accuracy in 12 clinically pathogenic microdeletion regions and determined that high FF is one of the most important factors for correct MD risk detection, followed by the observation, particularly in samples with lower FF, that longer microdeletion regions exhibit higher MD risk detection sensitivity.ConclusionsWe confirmed BinDel feasibility for fetal microdeletion risk detection in NIPT. Remarkably, the final BinDel tool correctly identified 88.2% (30 out of 34) MD cases, opening the possibility to integrate microdeletion analysis successfully into routine NIPT protocol. Additionally, we demonstrated that high FF is one of the most important factors for correct microdeletion risk estimation and that longer microdeletion regions display higher MD calling sensitivity. This work stands as a unique contribution to prenatal microdeletion screening, exhibiting a novel software simultaneously validated with a large microdeletion sample set, positioning it as the first of its kind in the field. BinDel is available athttps://github.com/seqinfo/BinDel.