The proband is a 5-year-old girl presented with conjugated hyperbilirubinemia and infantile cholestasis with the onset of clinical manifestation of features at 10 months of age.

The study design and protocol were conducted in accordance with the guidelines of the American College of Medical Genetics and Genomics and was approved by the Ethical Review Committee of Dr Lal Pathlabs. Written informed consent has been taken from parents of the proband included in the study, and the parents have provided consent to publish the data.

The DNA was extracted from 2 ml of the peripheral blood using Qiagen DNA mini kit, as per the manufacturer’s instructions. The quantity and quality of the extracted genomic DNA were measured by NanoDrop-2000 Spectrophotometer. Approximately 100 ng of genomic DNA was used to construct exome library using Ion Ampliseq Exome RDY Panel kit (Thermo Fisher Scientific). The resulting DNA library was quantified with Qubit dsDNA HS (High Sensitivity) Assay Kit on Qubit 3.0 Fluorometer. Approximately 25 pm of the library was used with the Ion Chef Instrument (Thermo Fisher Scientific) for template generation followed by enrichment of the templated ion sphere particles. Sequencing was performed using Hi-Q chemistry on Ion Proton system (Thermo Fisher Scientific).

The sequences were aligned against the reference genome (GRCh37/hg19) in Torrent Suite v.5.12.0 and Torrent Suite Variant Caller v.5.2.1 software (Thermo Fisher Scientific) with default parameters. The coverage analysis plugin and variant caller plugin from Life Technologies (Thermo Fisher) were used to analyze the Ion Proton sequencing run. Variant discovery, genotype calling of multiallelic substitutions, and indels were performed on each individual sample using the Torrent Variant Caller (TVC, version 4.6.0.7; Thermo Fisher). Statistics and graphs describing the level of sequence coverage produced for targeted genomic regions were provided by the Torrent Coverage Analysis (version 4.6.0.3). The variants were annotated by the Annotate variants 5.0 of Ion Reporter (Thermo Fisher).

Variants that were detected in the exome sequencing were filtered based on coverage (≥15x), minor allele frequency (≤0.01), and deleterious potential. All resulting variants were contrasted with the Human Gene Mutation Database [28] and Uniprot [29]. Furthermore, intronic, up- or downstream, and synonymous variants were removed. The pathogenicity of the detected variant was evaluated using Mutation Taster [30] and MutPredLOF [31]. Additional factors that were considered include the following: (1) absence in the general population; (2) novel appearance and disease phenotype from the family pedigree; (3) absence of any other mutation in JAG1 that could be responsible for the clinical phenotype; and (4) previous independent occurrence in an unrelated patient. An Integrative Genome Viewer [32] was used to visualize sequencing data. Variant frequencies were obtained from various databases such as the 1000 Genomes Project, dbSNP142, Exome Aggregation Consortium (ExAC) and gnomAD. Finally, for the interpretation of variant, American College of Medical Genetics and Genomics 2015 guidelines were used [33].

Confirmation of the mutation was performed by conventional Sanger sequencing using the BigDye Terminator v3.1 Cycle Sequencing kit (Applied Biosystems, Thermo Fisher Scientific) and was loaded on an ABI 3500Dx automated Genetic Analyzer (Applied Biosystems, Thermo Fisher Scientific). Primer sequences for the identified variant were designed using Primer 3.0 online as follows: Frd 5′- CCTCATTATTCGATGGCAAGGC -3′ and Rev 5′- GTTCTGTTCTTCAGAGGCCG-3′.

We detected a total of 39,679 variants comprising 55% (n=21,823) synonymous, 43% (n=17,062) missense, and 2% (n=794) frameshift or indel variants. For WES data filtering procedures, a filtering tree illustrated the step-by-step narrowing down of candidate gene or variants detected during next-generation sequencing data analysis (Figure 1). The first phase consisted of benign and synonymous variant filtering, and the second phase was based on variant impact (nonsynonymous and truncating), allele frequency (<0.1%), and pathogenicity prediction tools for missense variants (score >3). Since there were still a high number of candidate variants and genes, a second round of prioritization based on manual curation of biological function was performed, and variants in genes unrelated to ALGS were filtered out.

WES results indicated a novel heterozygous frameshift variant (c.3080delC;p.P1027RfsTer9) in the JAG1 gene responsible for ALGS (Figure 2A). The sequence alignment of heterozygous deletion (c.3080delC) at position Chr20:10621549 in JAG1 gene was viewed using the Integrative Genomics Viewer (Figure 2B). Sanger sequencing analysis confirmed that the proband carried the mutation in a heterozygous state (Figure 2C). This mutation is conserved across different species and can greatly affect the amino acid sequence of JAG1 gene that might change the protein function. Different web-based bioinformatics tools were used to analyze the pathogenicity of the variant and predicted this mutation to have potential damaging effects (Table 1).

ALGS is a highly variable autosomal dominant disorder, which involves multiple organ systems, and it requires a multidisciplinary team of medical specialists for its management [33-35]. The spectrum of mutations in JAG1 gene associated with ALGS includes full gene deletion and other protein-truncating mutations including nonsense, frameshift, and splice site as well as missense mutations, suggesting that the clinical phenotype is caused by haploinsufficiency for the JAG1 protein [9,15]. Phenotypic effects of JAG1 mutations are highly penetrant but with variable expressivity [36]. There is no strong correlation between the type and location of the JAG1 mutation and the severity of the disease, suggesting that other genomic modifiers beyond the known JAG1 mutation may be the cause of the variable expressivity that characterizes this disorder [9]. Unfortunately, no genotype-phenotype correlation exists between clinical manifestations and the specific JAG1 pathogenic variant or the location of the mutation within gene [37]. Although genetics of ALGS is well-defined, there is variable expressivity of the disease. Individuals with the same mutations, including patients belonging to the same family, show discordance in the phenotype [38]. In support of this concept, the genotype-phenotype correlation studies did not identify a link between the mutation type and clinical manifestation or severity [39].

To the best of our knowledge, we report here a novel variant underlining a frameshift mutation in the JAG1 gene using the Ion Torrent platform. In the patient currently under study, the onset of ALGS was at an early age with hyperbilirubinemia and infantile cholestasis. WES analysis revealed the previously unreported heterozygous c.3080delC variant in exon 25, which produces a truncated JAGGED1 protein due to a stop codon (p.P1027RfsTer9) and probably a diminished function of the Notch signaling pathway. A study reported that some symptoms associated with ALGS are deemed indicators of a bad prognosis, such as high total bilirubin levels between 12 and 24 months of age, liver fibrosis, and xanthomata [40]. The present patient had one of these predictors, high total bilirubin. However, further studies could clarify the effect of this mutation on the protein and signaling level. This JAG1 (c.3080delC) mutation has never been reported in public human databases, including the following: ClinVar [41], COSMIC [42], the 1000 Genomes Project [43], gnomAD [44], ExAC [45], dbSNP [46], and the Human Gene Mutation Database [28]. It was also not found in our in-house database of 1000 exomes (personal data). In Figure 3A and B, we illustrated all JAG1 pathogenic, or likely pathogenic, mutations reported in the public version of ClinVar and COSMIC databases according to ExAC frequency. These variants encompass frameshift (nucleotide‐level deletions, insertions, and insertion-deletions), nonsense (substitutions, start loss, and stop gain), missense, splice site, and in-frame deletions. We observed that the effect of mutations on phenotype and its severity differs between patients regardless of their mode of inheritance, which makes the understanding of the physiopathological mechanisms so far unknown. Interestingly, in the same exon 25, a missense mutation c.3080C>A;p.Pro1027Gln was previously reported in the COSMIC database (Figure 3B) which was disease-causing according to MutationTaster and damaging according to other in silico tools; FATHMM [47], MutPred [31], LRT [48], and EIGEN PC [49]. A 2019 study by Gilbert et al [50] showed that 94.3% of individuals with clinically diagnosed ALGS have a pathogenic variant in the JAG1 gene, 2.5% have a pathogenic variant in the NOTCH2 gene, and 3.2% are molecularly uncharacterized [50]. The spectrum of JAG1 mutations includes more frequently protein-truncating mutations (75%) and nonprotein truncating mutations (25%) [38,51].

In summary, our report broadens the spectrum of mutations in the JAG1 gene. It thus aids the geneticist for better identification of the etiological mutations leading to ALGS, thereby providing further insights to detect the associated hot spots, exons or mutations.

WES has revolutionized molecular genetic research and has become an essential genetic tool for molecular diagnosis of heterogeneous disorders. This study is an attempt to improve our understanding of the origin of ALGS caused by the identification of a variant in the JAG1 gene. The novel mutation identified here provides an appropriate course of management to the patient to offer genetic counseling to the family; it also offers to expand the genetic spectrum of JAG1-related ALGS, raise awareness among pediatricians on the morbidity of this severe form of ALGS, which is thus far underdiagnosed, and show them the added value of next-generation sequencing technology in reducing diagnostic wandering.