Common variants increase risk for congenital diaphragmatic hernia within the context of de novo variants

Abstract

Congenital diaphragmatic hernia (CDH) is a severe congenital anomaly often accompanied by other structural anomalies and/or neurobehavioral manifestations. Rare de novo protein-coding variants and copy-number variations contribute to CDH in the population. However, most individuals with CDH remain genetically undiagnosed. Here, we perform integrated de novo and common-variant analyses using 1,469 CDH individuals, including 1,064 child-parent trios and 6,133 ancestry-matched, unaffected controls for the genome-wide association study. We identify candidate CDH variants in 15 genes, including eight novel genes, through deleterious de novo variants. We further identify two genomic loci contributing to CDH risk through common variants with similar effect sizes among Europeans and Latinx. Both loci are in putative transcriptional regulatory regions of developmental patterning genes. Estimated heritability in common variants is ∼19%. Strikingly, there is no significant difference in estimated polygenic risk scores between isolated and complex CDH or between individuals harboring deleterious de novo variants and individuals without these variants. The data support a polygenic model as part of the CDH genetic architecture.

Keywords: WNT5A; common variants; congenital diaphragmatic hernia; de novo variants; genome-wide association study; single-nucleotide polymorphism.

Figure 1

Genome-wide association results in CDH (A) Manhattan plot of SNP-based association results in the discovery cohort. The x axis shows genomic coordinates, and the y axis shows the −log₁₀(p value) obtained by logistic regression adjusting for sex and the first four principal components. (B and C) Regional association plots of the 3p14.3 (B) and 7q36.3 (C) loci with surrounding genes. The y axis shows the −log₁₀(p value). Round points indicate genotyped SNPs, and star points represent imputed SNPs, colored by degree of linkage disequilibrium with the top association SNPs (purple points): rs55705711at 3p14.3 (B) and rs7777647 at 7q36.3 (C). (D–F) Forest plots demonstrating replication of the lead SNP at 3p14.3 rs55705711 (D), a second SNP at 3p14.3 rs3907611 (E), and the lead SNP at 7q36.3 rs7777647 (F) in an independent European cohort and the combined cohort. The x axis shows odds ratios (ORs) and 95% confidence intervals (CIs). n, sample size; AF, allele frequency of the risk allele.

Figure 2

Functional annotation of genome-wide-significant CDH loci (A and B) Alignment of functional data at 3p14.3 (A) and 7q36.3 (B). From top to bottom: Hi-C data derived from H1 human embryonic stem cell line (H1-hESC); TADs derived from H1-hESC; capture Hi-C data derived from H1-hESC; RefSeq genes; CDH GWAS SNP data from the combined cohort shown on a −log₁₀(p value) scale; transcription regulators from ReMap ChIP-seq in heart, brain, lung, and stem-cell-related-tissue track; single-nucleus ATAC-sequencing (snATAC-seq) data from normal human lung (A) and the 900 kb murine Shh distal enhancer containing multiple small enhancer regions with tissue specificity (B).

Figure 3

Polygenic risk scores in different subsets of the CDH testing cohort (A) PRSs in participants with CDH-affected individuals vs. control individuals, individuals with isolated vs. complex CDH, male vs. female individuals, or individuals with vs. without de novo damaging variants (D-DNV). p values were determined with a Student’s t test. (B) Polygenic risk for CDH in CDH-affected children (probands) compared to mid-parents (i.e., the average of mother and father risk scores) or control individuals. p values were determined with a Student’s t test. (C) Odds ratio of case-control proportions for PRSs at multiple thresholds. Exact p values by Fisher’s exact test are listed in Table S12.