Increased burden of de novo predicted deleterious variants in complex congenital diaphragmatic hernia

Abstract

Congenital diaphragmatic hernia (CDH) is a serious birth defect that accounts for 8% of all major birth anomalies. Approximately 40% of cases occur in association with other anomalies. As sporadic complex CDH likely has a significant impact on reproductive fitness, we hypothesized that de novo variants would account for the etiology in a significant fraction of cases. We performed exome sequencing in 39 CDH trios and compared the frequency of de novo variants with 787 unaffected controls from the Simons Simplex Collection. We found no significant difference in overall frequency of de novo variants between cases and controls. However, among genes that are highly expressed during diaphragm development, there was a significant burden of likely gene disrupting (LGD) and predicted deleterious missense variants in cases (fold enrichment = 3.2, P-value = 0.003), and these genes are more likely to be haploinsufficient (P-value = 0.01) than the ones with benign missense or synonymous de novo variants in cases. After accounting for the frequency of de novo variants in the control population, we estimate that 15% of sporadic complex CDH patients are attributable to de novo LGD or deleterious missense variants. We identified several genes with predicted deleterious de novo variants that fall into common categories of genes related to transcription factors and cell migration that we believe are related to the pathogenesis of CDH. These data provide supportive evidence for novel genes in the pathogenesis of CDH associated with other anomalies and suggest that de novo variants play a significant role in complex CDH cases.

Figure 1.

Fold enrichment of de novo variants in cases versus controls. The graph shows the fold enrichment of de novo variants in different variant categories for all variants, and variants in genes that were highly expressed in MDD. The fold enrichment is the ratio of variants in cases to variants in controls divided by the ratio of cases to controls. *P ≤ 0.05 and **P ≤ 0.01, as assessed by a binomial exact test.

Figure 2.

Haploinsufficiency in genes with different types of de novo variants in CDH cases. The boxplot shows the genes with likely gene damaging (LGD) de novo variants (nonsense, splicing and frameshift indels) or damaging missense (D-miss) de novo variants in CDH cases have higher probability of haploinsufficiency than the ones with synonymous (syn) or benign missense (B-miss) variants. P-value was calculated by Mann–Whitney U-test. The beeswarm plot embedded in the boxplot was made by R package ‘beeswarm’.

Figure 3.

ToppGene ranking in cases and controls. The graph shows the ranking of genes by ToppGene for (A) likely gene damaging (LGD) variants (nonsense, splicing and frameshift indels) and (B) likely deleterious (LGD plus deleterious non-synonymous). P-values are the result of a Mann–Whitney U-test.