Abstracts

Rationale:
The epilepsies are defined by their seizures, co-morbidities, electroencephalographic features and etiologies. Beyond the significant minority of people with epilepsy in whom a disease-causing variant can be identified, genetic factors are thought to play a much wider role in the epilepsies. People with common epilepsies are more likely than those without to carry ultra-rare genetic variants and these variants often affect genes already established in the causation of rare epilepsies. However, the effect of these variants on the clinical manifestation of the epilepsy is not known. We hypothesized that individuals carrying ultra-rare variants in the same gene have more similar epilepsies than would be expected by chance.
Method:
Data were collected in a standardized format from 37 international sites in years 1–2 of the Epi25 Collaborative study. We translated the clinical features into Human Phenotype Ontology (HPO) terms for analysis. The HPO models clinical knowledge about the relationships between phenotypic features according to hierarchies of specificity. For example, whilst a focal impaired awareness seizure and a focal aware seizure are distinct, both are types of focal seizure. With data in HPO format we quantified the clinical similarity of each possible pair of individuals using semantic similarity analysis, a technique based on information theory. We grouped individuals according to the presence of ultra-rare variants in 64 established epilepsy genes, compared their similarity scores to those expected from this cohort due to chance and identified the phenotypic features distinguishing those genes that reached statistical significance.
Results:
In a cohort of 6,023 individuals we found 3 genes in which ultra-rare predicted protein-altering variants, regardless of variant class, were associated with significantly similar phenotypes after correction for multiple testing, specifying a false discovery rate of 5%. These were STX1B, SLC6A1 and CACNA1E. STX1B (n = 4) was associated with atonic and myoclonic seizures, motor developmental delay and generalized epileptiform discharges; SLC6A1 (n = 14) with absence seizures, intellectual disability and generalized epileptiform discharges; and CACNA1E (n = 22) with focal epileptiform discharges. Analysis by the predicted consequences of variants suggested that whilst clinical similarity was greatest amongst those with severe variants, benign variants may also influence phenotype.
Conclusion:
In many epilepsies, individuals with ultra-rare variants in the same gene show remarkably similar clinical features. The phenotypic similarity extends to individuals with variants beyond those considered diagnostic in a clinical setting, suggesting as yet hidden genotype-phenotype associations in the epilepsies. Our results indicate that our method could help unravel the complex genetic architecture of the common epilepsies by identifying genes in which variants influence phenotype.
Funding:
:DLS is funded by a Wellcome Trust 4ward North Clinical PhD Fellowship (203914/Z/16/Z). Centers for Common Disease Genomics-funded Epi25 research activities at the Broad Institute, including genomic data generation in the Broad Genomics Platform, are supported by NHGRI grant UM1 HG008895 (PIs: Eric Lander, Stacey Gabriel, Mark Daly, Sekar Kathiresan). The Genome Sequencing Program efforts were also supported by NHGRI grant 5U01HG009088-02. Supplemental Epi25 phenotyping was supported by “Epi25 Clinical Phenotyping R03” NIH 1R03NS108145-01 (PIs: Daniel Lowenstein, Samuel Berkovic). Funders of Epi25 Collaborators contributing data are listed in the supplement to Epi25 Collaborative. Ultra-Rare Genetic Variation in the Epilepsies. Am. J. Hum. Genet. 2019;105:267–282.