EP388: Optical genome mapping capability expanded to enable detection of absence of heterozygosity

Abstract

Optical genome mapping (OGM) is a technique able to detect all classes of structural variants (SVs), including deletions, insertions, duplications, inversions, chromosomal aneuploidies, and translocations, as well as complex rearrangements across the whole genome that are undetectable by traditional methods such as genome sequencing and cytogenetics. OGM is now also able to detect copy-neutral absence of heterozygosity (AOH), which has traditionally been identified using chromosomal microarray and microsatellite analysis. Copy neutral absence of heterozygosity refers to a specific type of genetic condition where both alleles are the same and this can be a result of uniparental disomy or consanguinity. One consequence of AOH can be an increased susceptibility to recessive disease. Here we describe a method for AOH analysis based on optical genome mapping from the Bionano Genomics Saphyr system. Regions of AOH are identified by a consistent decrease in heterozygous SV calls across a genomic region in a case sample compared to the level observed in controls. DNA was processed using the Bionano Genomics Saphyr system and analyzed using the Bionano Access software. After filtering SV calls for high-quality, informative sites, AOH events were simulated by splicing together SV calling datasets from 153 controls and 4 double haploid samples, where AOH events were represented by regions derived from double haploid genomes, or regions derived from chromosome X in males. AOH regions were determined using a Hidden Markov Model (HMM), which was used to model the spatial dependence between neighboring SVs of a given zygosity: the hidden state was whether an SV belongs to an AOH or background region, and the observable states were whether the SV was called as homozygous or heterozygous. Model parameters were estimated by fitting the model to the simulated dataset, and performance was evaluated on newly simulated samples and additional samples with known AOH events that were previously identified using chromosomal microarray analysis. In simulated data, our method achieved high sensitivity and precision in detecting large AOH regions above 25 Mbp, with 92% sensitivity and 97% precision. Using a small cohort of samples with known AOH events, we have confirmed the utility of this method and were able to detect constitutional AOH events. In the current implementation, AOH events are detected if present in all cells; lower variant allele frequency events are not assessed. Our results show that it is possible to detect AOH regions using optical genome mapping alone. The performance for AOH detection by OGM is sufficient to detect the most relevant AOH events associated with constitutional disorders and adds an important new dimension to OGM technology. The new AOH algorithm is now available as part of Bionano Access version 1.7.