Genomic rare variant mechanisms for congenital cardiac laterality defect: A digenic model approach

Abstract

Laterality defects are defined by perturbations in the usual left-right asymmetry of organs. The genetic etiology that underlies congenital heart disease (CHD) is often unknown (less than 40%), so we used a digenic model approach for the identification of contributing variants in known laterality-defect-associated genes (n = 115) in the exome/genome sequencing (ES/GS) data from individuals with clinically diagnosed laterality defects. The unsolved ES/GS data were analyzed from three CHD cohorts: Baylor College of Medicine-Genomics Research to Elucidate the Genetics of Rare Diseases (BCM-GREGoR; n= 251 proband ES), Gabriella Miller Kids First Pediatric Research Program (Kids First; n = 158 trio GS), and Pediatric Cardiac Genomics Consortium (PCGC; n = 163 trio ES). trans-heterozygous digenic variants were identified in 2.8% (inherited digenic variants in 0.4%), 8.2%, and 13.5% of individuals, respectively; this was significantly higher than in 602 control trios provided by the 1000 Genomes Project (p = 0.001, 1.4e-07, and 8.9e-13, respectively). trans-heterozygous digenic variants were also identified in 0.4% and 1.4% of individuals with non-laterality CHD in Kids First and PCGC datasets, respectively, which was not statistically significant as compared to control trios (p = 1 and 0.059, respectively). Altogether, in laterality cohorts, 23% of digenic pairs were in the same structural complex of motile cilia. Out of 39 unique digenic pairs in laterality CHD, 29 are more likely to be potential digenic hits as predicted by the DiGePred tool. These findings provide further evidence that digenic epistatic interactions can contribute to the complex genetics of laterality defects.

Keywords: cilia; digenic; laterality defect; trans-heterozygous.

Figure 1

Conceptual framework of the study Individuals with laterality and non-laterality defects were analyzed for digenic variants in known 115 laterality genes from three cohorts, BCM-GREGoR, Kids First, and PCGC. A healthy population of 602 from 1000G were used as control cohort. Our analysis shows that the proportion of individuals carrying trans-heterozygous digenic variants differs significantly between control cohorts and laterality-defect cohorts (p = 0.001, 1.4e−07, and 8.9e−13 in BCM-GREGoR, Kids First, and PCGC vs. 1000G, respectively). However, no significant difference was observed between controls and non-laterality individuals (p = 1, 0.059 in Kids First and PCGC vs. 1000G). 29 out of 39 unique gene pairs identified in three laterality cohorts were likely to be potential digenic hits based on their pathway similarity, phenotype similarity, co-expression rank, protein-protein interaction distance, and pathway distance (p = 2.2e−16). Ten gene pairs are in the same structural complex of motile cilia. ES, exome sequencing; GS, genome sequencing. Out of those 29 potential digenic hits, the proportion of individuals carrying trans-heterozygous digenic variants differs significantly between control and laterality defect cohorts (p = 0.0004, 7.3e−06, and 4.5e−10 in BCM-GREGoR, Kids First, and PCGC vs. 1000G, respectively). However, no significant difference was observed between control subjects and non-laterality individuals (p = 0.609, 1 in Kids First and PCGC vs. 1000G).

Figure 2

Percentage of trans-heterozygous digenic variants identified in different cohorts trans-heterozygous digenic variants were identified in (A) 0.3% of control population (2 of 602 trio GS data from the 1000 Genomes Project), (B) 2.8% (7/251) of laterality CHD individuals and 0.4% (1/251) of laterality CHD individuals in the BCM-GREGoR cohort, who inherited both variants from affected father, (C) 2.1% of individuals in all CHD types in Kids First, (D) 8.2% of the laterality CHD individuals in the Kids First cohort, (E) 4.1% individuals in all CHD types, and (F) 13.5% of the laterality individuals in the PCGC cohort.

Figure 3

Biological functional annotation of the identified genes with trans-heterozygous digenic variants in three laterality cohorts Our analysis revealed that they were most significantly enriched in cilium movement, axoneme assembly, cilium-dependent cell motility, cilium assembly, axoneme dynein complex assembly, and cellular component assembly, with their −log(p) shown on the x axis.

Figure 4

Proportion of trans-heterozygous digenic variants in ciliary genes as compared to non-ciliary genes In three laterality cohorts, the proportion of digenic variants present in ciliary genes (27/55) is significantly higher than in non-ciliary genes (9/60), as indicated by p = 0.002.

Figure 5

Structural and functional features of identified ciliary and non-ciliary genes in probands Structure of motile cilia (A) (modified from Reiter and Leroux³⁴) and structure of non-motile cilia (B) (modified from Reiter and Leroux³⁴); gene pairs in gray boxes (C) represent those that were identified to carry trans-heterozygous digenic variants in laterality individuals in three cohorts, while gene pairs in green boxes represent those identified to carry trans-heterozygous digenic variants in non-laterality individuals in three cohorts.

Figure 6

A significantly higher proportion of the genes with the identified trans-heterozygous digenic hits in three cohorts are predicted to be a digenic gene pair (A) Using the previously described threshold (dashed lines), 218 (3.3%) out of the total of 6,555 potential gene pairs available from the 115 known laterality-defect genes are predicted to be digenic hits. (B) A significantly higher proportion, 74.3%, i.e., 29 out of 39 unique gene pairs identified in three laterality cohorts, are predicted to be digenic hits (p = 2.2e−16) compared to the total of 6,555 potential gene pairs available from the 115 known laterality-defect genes.