Long-Read Sequencing Resolves Complex Structural Variants and Identifies Missing Disease-Causing Variants in Unsolved Cases of Hemophilia

Introduction: Despite recent advances in genetic testing, many individuals with inherited bleeding disorders remain unsolved after a complete clinical evaluation. In some cases, no variant is found in a particular gene or genes of interest, while in other cases a large structural variant (SV) is found or suspected that cannot be resolved using existing methods. In the large U.S. hemophilia genotyping initiative MyLifeOurFuture, ~2% of participants with hemophilia A had no F8 variant identified despite exhaustive investigation. Additionally, large SVs were common in severe hemophilia and associated with inhibitor risk but were incompletely characterized in nearly all cases. We hypothesized that long-read sequencing (LRS) would identify missing disease-causing variants and fully characterize SVs associated with hemophilia in these individuals. Methods: We performed targeted or whole-genome LRS on the Oxford Nanopore platform of 25 hemophilia cases with FVIII levels <14% in which prior testing identified no pathogenic variant or identified a suspected SV impacting F8. After basecalling with guppy and alignment to GRCh38 using minimap2, single nucleotide and insertion or deletion variants were identified with Clair3. All variant calls were refined using population frequency and pathogenicity predictions. SVs were called with Sniffles2, SVIM, and CuteSV. We performed de novo assembly of SVs and called methylation in all samples. Methylation differences were identified using an in-house developed pipeline. Results: We identified candidate pathogenic variants and previously undetected SVs in 10/10 cases with no variant found on clinical testing. Candidate intronic variants predicted to affect splicing were validated by exon trap assays. SVs identified included deletions and large inversions. Among 11 cases with suspected SVs, LRS resolved the candidate SV providing detailed genomic structure, orientation, and predicted impact of the variant. Whole genome LRS resolved SVs involving the large segmental duplication within F8 intron 22 (Int22h1), revealing variants that have not previously been observed including unsuspected orientations in the F8 intron 22 positive control sample. We analyzed four females with unexplained low FVIII levels and identified skewed methylation profiles that may help explain their phenotype. Conclusion: In a single genetic test LRS can identify disease-causing variants missed by prior genetic testing, resolve complex SVs, and identify methylation differences that explain hemophilia phenotypes. These comprehensive methods are necessary to fully understand causes of hemophilia and complications such as inhibitor development. We expect this approach will similarly resolve "missing” genetic variation in other inherited disorders. Supported and funded by the BBI at UW, WACBD, HTRS MCRA (Genentech), and NHLBI TOPMed.