The Association Between Sleep Phenotypes and Epilepsy Genes

Abstract

Objectives: Sleep and epilepsy have a complex and bidirectional relationship. We aimed to uncover genetic mechanisms underlying this connection.

Methods: We leveraged genome-wide association studies that link sleep phenotypes and genetic loci and compared epilepsy genes with a random set of genes. Pathway analysis was applied to genes associated with both epilepsy and sleep phenotypes. To determine the specificity of our findings, we compared genes from other neurologic diseases and CNS cell types.

Results: A total of 8.1% of epilepsy genes (26/320) were associated with sleep phenotypes, 2.7-fold the rate of a random gene set (36/1,200 = 3%). Sleep-associated epilepsy genes were enriched in biological processes involving brain development and neuronal function. Other CNS disease genes also demonstrated strong genetic links to sleep phenotypes while peripheral nervous system diseases genes did not. Genes expressed in neurons, astrocytes, and oligodendrocytes were highly associated with sleep phenotypes while those in microglia and endothelial cells were not.

Discussion: While epilepsy and sleep phenotypes share genetic links, this finding extends to other CNS diseases and genes expressed in brain cells. The brain's effect on sleep likely has a genetic underpinning, as variation in genes expressed in cells responsible for brain function affects sleep phenotypes.

Figure. Comparison of the Smallest p Values (MIN-P) for Each Gene Set vs Those for Random Genes

For all single-nucleotide variants (SNVs) associated with each gene in a gene set, the sleep GWAS data set was queried to extract p values, which indicate how significantly a SNV is associated with a sleep phenotype. The smallest p value for each gene was designated as MIN-P. From the MIN-P values for all the genes in a gene set, the median was determined, subject to a negative log₁₀ transformation, and plotted along the x-axis, with higher values showing stronger association of the gene set with sleep phenotypes. For comparison, this value for a random set of 1,200 genes was 3.6 (vertical dotted line). For the y-axis, the distribution of MIN-P values for each gene set was compared with those of a random gene set, using a Wilcoxon rank-sum test, and the negative log₁₀ transformation of the Bonferroni-adjusted p-value for this comparison was plotted, with higher values carrying higher statistical significance. The horizontal dotted line designates a threshold of significance of p = 0.05. Genes expressed in neurons had the largest degree of difference in MIN-P values when compared with random genes. GWAS = genome-wide association study.