Variant segregation in rare disease singletons and duos using nanopore adaptive sampling

Determining the pathogenicity of candidate variants in patients with rare disorders often involves segregating candidate variants to identify whether they occur in trans (recessive disorders) or determining if they are of de novo origin (dominant disorders). Variant segregation typically requires access to DNA from both parents. However, obtaining DNA samples from both parents is often not possible, leading to inconclusive genetic test results. Here we propose a cost-effective method that utilizes nanopore selective sequencing of the affected individual. In recessive genes of interest, two or more candidate variants can be phased using only patient DNA. By combining this approach with short-read data from a single parent, we can determine if a candidate variant is present on the allele contributed by that parent (50% of variants) and, consequently, if it originated from a de novo mutation. Using a panel of 271 clinically relevant genes, we predicted proband’s haplotype with the causative allele on their paternal or maternal haplotype. Predictions were verified with trio data. Our model exhibited a precision of 99.42%, a recall of 70.48%, and a false discovery rate below 1%. Importantly, all false positives were identified through manual inspection of the supporting reads and variants using a genome browser. Our method identified the parental haplotype of causal de novo allele in all patients involved in this study, which were unknown with trio data. Overall, our method represents an effective approach for segregating variants in individuals without strict access to parental DNA, addressing a critical need in clinical practice.