Heritability and genetic determinants of normal human facial shape and body size

Abstract

Human appearance is a complex assemblage of highly variable anatomic structures that together make each of us unique. Familial similarity, particularly in the face, underscores a genetic component; however, little is known about the genes that underlie differences in physical appearance. We estimated heritability of facial and anthropometric traits in African children and adolescents using genetic relatedness to jointly estimate variance explained by common genetic variation and uncaptured genetic variation, the sum representing narrow-sense heritability. Our facial trait estimates range from 17-67%, with horizontal measures being the most heritable. Facial traits measured in the same physical orientation have high positive genetic correlations, whereas traits in opposite orientations have high negative correlations. Among our anthropometric traits, head circumference and hand length measurements were the most heritable (64.9-98.2%) and were similar to previous European estimates, unlike our estimates for height, weight, and BMI (35.1-48.5%). We carried out GWAS and replication of facial traits in Africans, GWAS of facial traits in Europeans, and GWAS of anthropometric traits in Africans. In Africans, we identified genomewide significant association SNPs in SCHIP1 and PDE8A with facial size and suggestive association of ten additional loci with facial shape. In Europeans, we identified significant association with several different facial traits near the following genes known to be involved in the face: MAFB, PAX9, MIPOL1, ALX3, HDAC3, and PAX1. GWAS of anthropometric traits in Africans identified 49 loci with suggestive association. Candidate genes within these loci that have been previously associated in Europeans or demonstrate biologically relevant phenotypes in humans, mice, or cells include RCAN2, ZEB1, SYT1, CCDC91, PDE4D, PDZRN3, and LCT. Using anthropometric data from Africans, we developed normal growth curves and compared these to several references. We identified significant growth faltering that was associated with community-level, academic outcome, and environmental variance of facial shape. We demonstrate the importance of monitoring growth into adolescence using population-specific standards, and show that the environment plays a large role on growth. Overall we present multi-ethnic data on normal variation of facial shape and body size that augment our understanding of physical development and provide insights relevant to human disease and forensics.