Integrative approaches generate insights into the architecture of non-syndromic cleft lip with or without cleft palate

Abstract

Summary Non-syndromic cleft lip with or without cleft palate (nsCL/P) is a common congenital facial malformation with a multifactorial etiology. Genome-wide association studies (GWASs) have identified multiple genetic risk loci. However, functional interpretation of these loci is hampered by the underrepresentation in public resources of systematic functional maps representative of human embryonic facial development. To generate novel insights into the etiology of nsCL/P, we leveraged published GWAS data on nsCL/P as well as available chromatin modification and expression data on mid-facial development. Our analyses identified five novel risk loci, prioritized candidate target genes within associated regions, and highlighted distinct pathways. Furthermore, the results suggest the presence of distinct regulatory effects of nsCL/P risk variants throughout mid-facial development and shed light on its regulatory architecture. Our integrated data provide a platform to advance hypothesis-driven molecular investigations of nsCL/P and other human facial defects.

Prioritization of non-coding elements involved in non-syndromic cleft lip with/without cleft palate through genome-wide analysis of de novo mutations

“…The datasets included genome-wide maps of eight chromatin states from hNCCs, 24 cranial neural crest cells (cNCCs), 25 and human facial embryonic tissues, 26 which had been aggregated in a previous study by our group. 4 Furthermore, general genomic features with a priori evidence for functional relevance or evolution were included; i.e., (1) 4,307 evolutionarily highly conserved non-coding elements (CNEs) based on a prior publication, 27 and (2) 1,570 enhancer regions from the VISTA enhancer browser 28 ( supplemental methods ).…”

“…To detect local enrichments of non-coding DNMs independent of genomic features (comparable with gene-burden tests for protein-coding variants), DNMs were combined based on their location within regulatory units; i.e., topologically associating domains (TADs). Positional data were retrieved for 2,991 TADs from human embryonic stem cells, as described elsewhere, 4 and enrichment of DNMs in TADs was tested using FunciVar ( supplemental methods ). Given the considerable burden of multiple testing with regard to the present sample size, we additionally defined a set of 45 candidate TADs on the basis of recent GWAS results, as previously described 4 (TADs GWAS , Table S1 ).…”

“… 3 Recent genome-wide association studies (GWASs) have identified common risk variants at 45 genomic loci, which explain about 30% of phenotypic variance in Europeans. 4 Research suggests that further types of genetic variation may also contribute to disease risk, including variants from the low-frequency part of the allelic spectrum. For example, previous studies have identified private and rare risk variants for nsCL/P in genes underlying orofacial cleft syndromes within multiplex families, 5 in genes involved in epithelial cell adhesion processes, 6 and in genes located within GWAS loci.…”

Decoding the Human Face: Progress and Challenges in Understanding the Genetics of Craniofacial Morphology

“…These enrichments or frequencies for the set of GWAS loci can then be compared with enrichments from a background set of loci not associated with facial shape to estimate the significance of enrichment. Such an approach has been applied to both normal-range facial GWAS loci ( 25 ) and loci associated with nsCL/P ( 99 , 163 ) and has consistently found significant enrichment or overlap with active regulatory elements (mostly enhancers rather than promoters) in both in vitro–derived CNCCs and embryonic craniofacial tissue, as compared with other embryonic and adult human cell types and tissues that have been examined.…”

“…Perhaps the most straightforward is to generate an epigenomic map from a cell type or tissue thought to have an important role in the trait of interest, with the rationale that associated variants located in regulatory regions are active in the cell type of interest. This approach has been widely used by several GWASs of face shape and/or disease ( 25 , 99 , 163 , 164 ), due in part to its relative ease: Once epigenomic maps have been generated from sufficient cell numbers in the relevant ex vivo– or in vitro–derived cell type, they can be reused in future GWASs, and the fine mapping can be performed genome-wide. Stronger evidence can be provided by perturbing the activity of candidate regulatory elements harboring associated variants and assessing the impact on the expression of nearby genes.…”