Comprehensive analysis of coding variants highlights genetic complexity in developmental and epileptic encephalopathy

Abstract

Although there are many known Mendelian genes linked to epileptic or developmental and epileptic encephalopathy (EE/DEE), its genetic architecture is not fully explained. Here, we address this incompleteness by analyzing exomes of 743 EE/DEE cases and 2366 controls. We observe that damaging ultra-rare variants (dURVs) unique to an individual are significantly overrepresented in EE/DEE, both in known EE/DEE genes and the other non-EE/DEE genes. Importantly, enrichment of dURVs in non-EE/DEE genes is significant, even in the subset of cases with diagnostic dURVs (P = 0.000215), suggesting oligogenic contribution of non-EE/DEE gene dURVs. Gene-based analysis identifies exome-wide significant (P = 2.04 × 10^-6) enrichment of damaging de novo mutations in NF1, a gene primarily linked to neurofibromatosis, in infantile spasm. Together with accumulating evidence for roles of oligogenic or modifier variants in severe neurodevelopmental disorders, our results highlight genetic complexity in EE/DEE, and indicate that EE/DEE is not an aggregate of simple Mendelian disorders.

Fig. 1

Patterns of excess of URVs in EE/DEE. a Result of logistic regression analysis testing association between each type of URVs and the case−control status. Odds ratios for one additional URV and 95% confidence intervals are plotted (i.e. one additional URV changes the risk of being EE/DEE with the indicated odds ratio). Uncorrected P values for each test are shown beside the plots. Plots are color-coded as follows: null, red; Moderate (defined by SnpEff, e.g. missense and inflame), orange; synonymous, blue; noncoding, gray; and shape-coded as follows: overall functional type with no subclassification (e.g. null and Moderate), filled square; URVs more likely to be functional (i.e. consensus-damaging (CD) missense and frontal cortex DNase I hypersensitive site (FCDHS) and splice region (SR) synonymous), filled circle; URVs less likely to be functional, open circle. The numbers of URVs subjected to each analysis are indicated in the brackets. b, c Results of logistic regression analysis focusing on LOF-intolerant genes (probability of being LOF-intolerant (pLI) > 0.9) (b) and known 58EE/DEE genes (Supplementary Table 3) (c), respectively. Statistically significant results considering the total number of hypotheses tested in Figs. 1 and 2 (n = 66) are indicated as follows: **Bonferroni-corrected P < 0.05 (raw P < 0.000758), *Benjamini−Hochberg-corrected P < 0.05 (raw P < ~0.02)

Fig. 2

Excess of dURVs in non-58EE/DEE genes in individuals with or without pathogenic mutations. a Result of logistic regression analysis testing association between each type of URVs in genes not included in 58EE/DEE genes (non-58EE/DEE genes) and the case−control status. Odds ratios for one additional URV and 95% confidence intervals are plotted. Uncorrected P values for each test are shown beside the plots. Plots are color-coded as shown in the legend of Fig. 1. The combined group of damaging (null and CD missense) URVs (dURVs) was indicated by the purple diamond. b−d Results of analyses of URVs in non-58EE/DEE genes comparing control subjects with the following subsets of case group: cases not carrying dURVs in 58EE/DEE genes (b, n = 605), cases carrying dURVs in 58EE/DEE genes (c, n = 138), and cases carrying convincingly pathogenic URVs (pURVs; e.g. confirmed de novo mutations in 58EE/DEE genes; see the main text and Methods) (d, n = 116). e, f Results of analyses stratifying cases with pURVs in 58EE/DEE genes into subsets. In (e), groups of null pURV carriers (n = 45) and CD missense pURV carriers (n = 71) were analyzed for the burden of dURVs in non-58EE/DEE genes. In (f), we analyzed groups of individuals carrying; pURVs previously reported in non-EE/DEE phenotypes (n = 10, based on information in the Human Gene Mutation Database), pURVs previously reported in EE/DEE (n = 33), and the other pURVs with no or uncertain previous report (n = 73). In (a−f), genes of interest included in the analysis are indicated in the blue boxes (as in Fig. 1), and the case subsets contrasted with the controls are indicated in the light green boxes. The number of URVs analyzed in each test is indicated in the brackets. Statistically significant results considering the total number of hypotheses tested in Figs. 1 and 2 (n = 66) are indicated as follows: **Bonferroni-corrected P < 0.05 (raw P < 0.000758), *Benjamini−Hochberg-corrected P < 0.05 (raw P < ~0.02). g A diagram indicating relationship among the case subsets subjected to the analyses

Fig. 3

Gene-based burden test of dURVs. a Manhattan plot of P values for each gene obtained from the gene-based burden test. 58EE/DEE genes are indicated by the red-filled black circles. The red and blue horizontal lines indicate the thresholds for exome-wide significance (P = 2.5 × 10⁻⁶) and the nominal significance (P = 0.05), respectively. Genes with P < 0.005 are labeled with their gene symbols. b The observed overlap between genes with nominally significant (2.5 × 10⁻⁶ ≤ P < 0.05) dURV burden and 58EE/DEE genes (the red vertical line: 12.4% of the nominally significant genes overlapped with 58EE/DEE genes), and an expected distribution of overlaps obtained from random shuffling of the case−control labels in our cohort (cyan area). The corresponding P value calculated from 10,000 times of random shuffling is shown above the red vertical line. c Networks of gene ontology terms overrepresented among the non-58EE/DEE genes with nominally significant burden of dURVs in EE/DEE. Nodes are color-coded by clusters. Node size represents the significance of enrichment. Edge width represents the overlap coefficient. d A schematic representation of the exon−intron structure (adapted from ExAC Browser) of NF1 and the locations of damaging DNMs identified in our EE/DEE cohort

Fig. 4

Insights into the genetic architecture of EE/DEE. Summary of the insights into the genetic architecture of EE/DEE obtained from this study. Adapted from Fig. 1 of ref. . Findings in this study, corresponding spectrum of allele frequency/effect size and related figures/tables are indicated