Target capture next‐generation sequencing in non‐inversion haemophilia: an alternative approach

Haemophilia A (HA) and haemophilia B (HB) are two forms of an X-linked monogenic hereditary disorder caused by mutations in the coagulation factor VIII (F8) and factor IX (F9) genes respectively.1-3 The prevalence of the disorders is 1 in 5000 (HA) and 1 in 30 000 (HB) male births.4,5 Haemophilia is classified into three clinical phenotypes based on FVIII and FIX residual clotting activity: severe (<1%), moderate (1–5%) and mild (5–40%).6,7 The type of mutation is related to the phenotype. To date, 2582 and 1112 unique variants have been associated with the majority of HA (https://www.LOVD.nl/F8) and HB (https://www.LOVD.nl/F9) cases respectively. Determining the causative genetic variants in families affected by haemophilia is important for pregnancy, neonatal management and to inform about the risks of forming a neutralizing antibody and bleeding.2,8,9 Clinically, the most frequently used molecular method of diagnosing non-inversion haemophilia risk variants in clinical practice is Sanger sequencing. However, its limited throughput and low cost-effectiveness restrict simultaneous detection of the F8 (186 kb), F9 (34 kb) and VWF (178 kb) coding regions during a single workflow. Sanger-based sequencing in a clinical diagnostic laboratory is usually used to sequence the phenotypically indicated gene; additional potentially pathogenic variants that may contribute to the phenotype in such patients may not be examined, which can lead to an incorrect diagnosis and therapy. For example, low FVIII levels in patients may be caused by a VWF mutation rather than an F8 mutation.10 Therefore, it is important for clinicians to obtain more accurate and comprehensive information about haemophilia and diseases with similar clinical phenotypes. Next-generation sequencing (NGS) is an alternative state-of-the-art technology that promises simultaneous identification of multiple genes at a manageable cost while overcoming the deficiencies of Sanger sequencing, and differentiates phenotypic overlaps, which would significantly reduce the risks described above. To determine whether NGS technology can be used as an alternative method for detecting non-inversion haemophilia risk variants in routine clinical practice, we established our NGS method on the Ion Torrent platform and designed an AmpliSeq target capture panel including three candidate genes (F8, F9 and VWF) that are related to haemophilia and von Willebrand disease (vWD). All gene-coding sequences, untranslated regions (UTRs), and 10 bp exon-flanking intronic sequences, which probably determined the splicing sites, were designed to be easily amplified by multiplex PCR reaction using AmpliSeq technology (Thermo Fisher, Waltham, MA, USA). The experiment revealed high efficiency and specificity, with an average coverage of 98·68% and uniformity >90%. Very small missing gaps were mostly located in the UTRs or complex sequence structure regions and were complemented by extra Sanger sequencing. General data processing and bioinformatics analyses were performed as shown in Fig. 1. To assess the diagnostic effectiveness and feasibility of the established NGS platform for clinical applications, we analyzed 1 572 subjects from 239 HA pedigrees and 76 HB pedigrees. All cohorts were fully informed of the aim of the study and provided written informed consent. The data were compared to those of Sanger sequencing (Table I). The HA detection rate was 92·88% (222/239) for NGS and 90·37% (216/239) for Sanger sequencing. The NGS platform additionally detected three large F8 fragment deletions in three HA pedigrees and three VWF risk variants in another three pedigrees previously diagnosed as HA. The HB detection rate was 94·74% (72/76) for NGS and 89·47% (68/76) for Sanger sequencing, and the NGS platform additionally detected four large F9 fragment deletions in four HB pedigrees. NGS simultaneously detected F8, F9 and VWF for all subjects in one workflow. This is different from current clinical detection, which usually detects a single pathogenic gene and confirms the disease-causing variants found previously in the index patient for other family members, which might result in some incidental findings. We found 18 patients and 20 carriers with two or more risk variants described in one gene or two different genes (Table SI). Among these subjects, an additional 13 VWF risk variants were identified in patients, and five were novel variants (c.3197_3201del, c.C6796T, c.5664+1G>T, c.997+1G>C and c.T595C). We only detected risk variants in VWF from three pedigrees, but we did not find any pathogenic variants in F8 or F9 (Table SII). Similar cases have been reported in previous reports.11, 12 No potentially causative variants were found in another 21 haemophilia pedigrees (17 HA and 4 HB) in this study. We speculate that the genetic variants that adversely impact F8 or F9 gene function and cause haemophilia lie in genomic regions outside the target captured by our NGS and Sanger sequencing methods. In order to facilitate processing and analyzing NGS data independently for clinicians and testers, we developed an integrated variant analysis system to accompany the NGS platform. The system contains six modules: quality control, annotation, filtering, interpretation, reporting and statistics. Benign and likely benign variants are removed in the reporting module, and pathogenic and likely pathogenic variants are highlighted as the top candidate variants after manual correction and evaluation of the remaining variants. Variants of uncertain significance are also shown in the report to provide a more likely reference for the clinician. This integrated variant analysis system greatly reduces the number of candidate variants requiring consideration by the clinician and avoids missing harmful variants caused by manual analyses of the sequencing results. One of the most interesting and practical modules is the statistics module. The frequency of each variant detected in the laboratory can be counted with this statistics module. It is helpful to determine high-frequency haemophilia mutations in the Chinese population by gradually accumulating experimental data. In summary, based on data from 315 pedigrees previously diagnosed as haemophilia, the established NGS platform can be an alternative method of detecting non-inversion haemophilia risk variants in routine clinical practice. It can provide more accurate and comprehensive variant information to clinicians using less time and less manpower with the help of the integrated variant analysis system. Funding for the project was provided by the Science and Technology Program of Guangzhou (Grant No. 201604020104, 201704020114, 201803040009), the Science and Technology Program of Guangdong (Grant No. 2018A030313286) and the Medical Scientific Research Foundation of Guangdong (Grant No. A2017518). D-X L, L-Y L and X-X Y designed the research study; F-X L, W-J J and Y-H L collected clinical specimens; L-M H, X Y, Y-F X and D-M F performed the research; S L and B-W H analysed the date; Q L and J-J C wrote the paper. The authors have no competing interests. Table SI. Patients and carriers in whom two variants were detected (haemophilia). Table SII. Individuals previously classified as haemophilia patients/carriers with VWF risk variants. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.