Integrative Analysis of Exome Sequencing Identifies Novel Associated Genes in Craniosynostosis

Abstract

Background: Craniosynostosis (CRS) is a pathologic condition caused by premature ossification of the cranial sutures, leading to deformation of the skull and abnormal brain development. Although genetic factors play crucial roles in CRS, identifying a causal gene in most cases remains challenging. We aimed to identify novel CRS-associated genes in mostly unsolved CRS cases, based on the polygenic inheritance model. Methods: We included 95 Korean CRS cases with 42 trios. This cohort consisted of 82 unsolved cases having a previous targeted sequencing and 13 cases for initial molecular evaluation. The whole-exome sequencing was performed for 179 genomic DNA samples (95 probands, 84 parents). We generated an ancestry-matched control dataset (n = 4,595) by merging three genome sequencing databases (NARD, KRGDB, and Korea1K). Rare variant (minor allele frequencies < 1%) burden analysis was performed for autosomal haploinsufficient genes (pLI > 0.9; n = 2,813) to identify novel genes associated with CRS. The accumulative counts of rare variants among the identified CRS risk genes were compared between the CRS patients, their parents, and a healthy control population (Korea1K). In addition, pedigree analyses and gene network analyses were carried out to investigate genetic interactions between the identified CRS risk genes. Results: We identified 40 CRS risk genes (P < 0.05) of which 34 had relevant phenotypic associations such as craniofacial, neurodevelopmental, and bone phenotypes in humans and/or mice. Five probands had chromosomal abnormalities with no significant rare variant burden (0.60 variants/exome), while probands with no chromosomal abnormalities (n = 90) had significantly higher rare variant burden (2.47 variants/exome) compared with their parents (n = 84; 1.92 variants/exome, P = 0.010) and the unaffected control group (n = 1,048; 1.17 variants/exome, P < 0.001). Pedigree and gene network analysis suggested 10 gene pairs in 13 genes as digenic disease-causing combinations; 13 probands had at least one pathogenic pair, while their parents and the unaffected controls had none of these combinations. Conclusion: This study identifies novel CRS risk genes and extends the inheritance mode of CRS from monogenic to polygenic. Deleterious rare variant burden in CRS risk genes may collectively contribute to the disease phenotype. A polygenic cause of CRS should be considered for molecular evaluation, clinical interpretation, and genetic counseling, particularly among unsolved cases.