Mosaic Marfan syndrome patients identified

Most patients with Marfan syndrome (MFS), a connective-tissue disorder, harbor heterozygous variants in FBN1, a gene encoding fibrillin-1. Typically, individuals harboring mosaic pathogenic FBN1 variants have been reported to be asymptomatic. In this issue, Arnaud and colleagues identify five MFS probands with a mosaic pathogenic variant in FBN1. The findings indicate that bioinformatic analysis parameters for molecular diagnosis warrant an update, according to the authors. The researchers sequenced genomic DNA from blood samples of 5,000 candidate MFS patients. They screened the FBN1 gene either by bidirectional Sanger sequencing or, for more than 3,000 patients, by next-generation sequencing (NGS) using the MARFAN MASTR assay or a custom capture array. Designed to capture FBN1 as well as 27 other genes associated with MFS and related diseases, the custom NGS capture panel led to the identification of at least one heterozygous pathogenic FBN1 variant in about half of the nearly 2,000 probands shown to harbor variants in the gene. The panel also led to the identification of five mosaic pathogenic FBN1 variants in five different probands. The five identified mosaic probands were sporadic cases without children and with unaffected parents who did not carry the pathogenic variant. Four of the five patients presented with classic MFS clinical features, including ascending aortic dilation and/or ectopia lentis, and were diagnosed as infants. The last patient was diagnosed at 48 years old following emergency surgery for type A aortic dissection. The findings are unexpected, as mosaicism has rarely been associated with symptomatic MFS, let alone severe MFS presenting in infancy. The researchers recommend testing by NGS capture panel for patients with typical MFS presentation for whom no single-nucleotide variant or exon deletion/duplication is identified. —V. L. Dengler, News Editor

Using data mining to identify patients for molecular diagnosis

The range of phenotypic and genotypic heterogeneity in neurodevelopmental syndromes renders genetic diagnosis complicated. Patients with variants in the voltage-gated potassium channel subfamily member KCN2A present on a clinical spectrum from isolated intellectual disability to epileptic encephalopathies. Hully and colleagues used a data-mining system to scour a warehouse of electronic medical records for patients with phenotypes similar to those of an individual with a known variant. The researchers first recognized that two patients who presented with highly similar clinical features also carried the same de novo heterozygous variant in KCN2A. Due to the strong genotype–phenotype correlation in these patients, the researchers then searched for other patients who might harbor the variant. Using a data-mining system called Dr. Warehouse, the team searched 3 million clinical narrative reports, representing 500,000 patients, for others with clinical features similar to those of the first patient. The data mining identified five patients with the highest similarity indices. The patient with the highest similarity score shared symptoms with the original two patients, including early-onset hyperexcitability, myoclonic jerks, constant screaming, no eye contact, swallowing difficulties, and recurrent multifocal seizures. A targeted next-generation sequencing (NGS) panel for the high-similarity patient and his parents identified the same de novo heterozygous KCNA2 variant found in the previous patients. The remaining four identified patients had lower similarity scores. The NGS panel did not detect the pathogenic KCN2A variant in the three individuals who provided consent for genetic diagnostic testing. Further analysis extracted phenotypic key concepts from the original patient’s clinical reports over time, allowing a description of the disease’s natural progression. The authors conclude that such database mining could facilitate identifying patients with rare genetic conditions and predicting their natural history. —V. L. Dengler, News Editor