New Entries in the Lottery of Facial GWAS Discovery

We are all familiar with the strong genetic control of faces seen in the almost indistinguishable facial appearance of identical twins, similarities within families and populations, as well as shared facial characteristics in medical syndromes. Nevertheless, our understanding of the genetic architecture of normal-range human facial variation has remained largely uninvestigated, until recently. Cole et al., 2016 [1], and Shaffer et al., 2016 [2], present us with additional insights into the complex puzzle that is facial genetics. These two studies along with one other, Ahidkari et al., 2016 [3], represent the second round of genome-wide association studies (GWAS) on facial morphology; the first round appeared in 2012 [4,5]. Although genes affecting the development of the face have been studied for decades using pedigree-based analyses, molecular biology, and animal models, the discovery of genetic variants affecting normal-range facial variation using GWAS is still in its infancy. The progression of facial GWAS has been impeded by several factors, including a lack of appropriate 3-D facial scanning systems, the significant efforts required to collect participants, subsequent rigorous statistical analyses, and the lack of methods for functional follow-up analyses [6]. The two facial GWAS from 2012 provided a handful of associated genes, one of which, PAX3, was significant in both studies (Fig 1). Evidence confirming a role for PAX3 is presented in the Shaffer et al. report, which also replicated three other genes from the 2012 GWAS: CACNA2D3, C5orf50, and PRDM16 [2]. The failure to replicate the other associations could result from many factors in study design, such as consistent facial phenotyping as noted by [2] and population background as noted by [1]. Future studies bear the responsibility to investigate earlier findings for replication, and with these two new facial GWAS, this list has tripled in length. These two GWAS provide some new and exciting loci, datasets, and approaches. The increased list of craniofacial genes brings us closer to a better understanding of facial development and disease. They also increase our optimism in realizing unresolved medical and forensic applications that have compelling social implications. Although many genetic conditions that involve clinically significant (atypical) patterns of facial development have been mapped, and these studies have arguably established most of what we know about facial genetics, many individual cases remain undiagnosed. A better understanding of the genetics of normal-range craniofacial development can and should help in the delineation of which genes underlie these conditions. Generally, this process has been inverted, with information on the genetic determinants of disorders serving to help identify candidate genes for investigations into normal-range facial variation [7]. Understanding the genetic architecture of normal-range facial variation would also assist efforts to predict faces from DNA [8], a forensic application that has generated much public interest.