Abstract
The human face is a highly complex and variable structure resulting from the intricate coordination of numerous genetic and non-genetic factors. Hundreds of genomic loci impacting quantitative facial features have been identified. While these associations have been shown to influence morphology by altering the mean size and shape of facial measures, their effect on trait variance remains unclear. We conducted a genome-wide association analysis for the variance of 20 quantitative facial measurements in 2,447 European individuals and identified several suggestive variance quantitative trait loci (vQTLs). These vQTLs guided us to conduct an efficient search for gene-by-gene (G × G) interactions, which uncovered an interaction between PRICKLE1 and FOCAD affecting cranial base width. We replicated this G × G interaction signal at the locus level in an additional 5,128 Korean individuals. We used the hypomorphic Prickle1Beetlejuice (Prickle1Bj) mouse line to directly test the function of Prickle1 on the cranial base and observed wider cranial bases in Prickle1Bj/Bj. Importantly, we observed that the Prickle1 and Focadhesin proteins co-localize in murine cranial base chondrocytes, and this co-localization is abnormal in the Prickle1Bj/Bj mutants. Taken together, our findings uncovered a novel G × G interaction effect in humans with strong support from both epidemiological and molecular studies. These results highlight the potential of studying measures of phenotypic variability in gene mapping studies of facial morphology.
Highlights
Human genetic studies have been remarkably successful at detecting variants with an impact on mean trait values
We discovered a novel G × G interaction effect between PRICKLE1 and FOCAD impacting cranial base width, which was statistically replicated in an independent cohort and experimentally verified by showing the co-expression of both genes during the critical stages of mouse craniofacial development
We followed up on the human PRICKLE1 vQTL to explore the source of the phenotypic variance heterogeneity, with a particular focus on its G × G interaction effect
Summary
Human genetic studies have been remarkably successful at detecting variants with an impact on mean trait values. Variance quantitative trait loci (vQTLs) are genetic variants exhibiting inter-individual intragenotypic variability, where one of the alleles is associated with a larger phenotypic variance compared to the other (Rönnegård and Valdar, 2012). VQTLs are seldom examined in conventional genome-wide association studies (GWASes), which usually assume variance homogeneity across genotype groups and aim to detect differences in group means. Recognized as one of the critical potential sources of missing heritability (Zuk et al, 2012), G × G interactions remain largely uncharacterized in most human traits due to low statistical power, a high computational burden, and difficulties in uncovering the biological relevance for statistical interactions. As a genome-wide search for G × G interactions would be intractable, narrowing the focus to a subset of candidate SNPs is necessary, and effective strategies are needed to guide this pre-selection process
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
