A Preliminary Genetic Investigation of Normal Variation in Facial Features

Open Access
Matthes, Kerri Allison
Graduate Program:
Master of Science
Document Type:
Master Thesis
Date of Defense:
April 08, 2008
Committee Members:
  • Mark Shriver, Thesis Advisor
  • admixture mapping
  • craniofacial morphology
  • candidate genes
Variation in facial features between populations has arisen over time due to differing evolutionary forces acting on populations, resulting in allele frequency and phenotypic differences. Investigating variation in facial morphology can provide a more thorough understanding of these evolutionary forces and can also be used to identify genes involved in normal facial feature variation. This study aims to elucidate preliminary genetic relationships between ancestry, facial morphology and candidate genes that may play a role in morphological variation. Because admixed populations are well suited for studying traits that differ between two parental populations, a sample of African Americans were chosen for investigation in this study. The maximum likelihood genetic ancestry for 131 participants, ages 18-37, was obtained using 176 ancestry informative markers to determine their proportions of European and West African ancestry. Using 3dMDface (Atlanta, GA) imaging and analysis software, 3-dimensional images of each participant were captured and landmarked at 22 standard anthropometric locations. These coordinates were compared using a Euclidean Distance Matrix Analysis to identify morphological trait variations within the sample. DNA from each participant was genotyped at a polymorphic site in seven different selection-nominated candidate genes known to be involved in genetic disorders with craniofacial phenotypes. WinEDMA Form comparison between West African and European pseudo-parental groups of the same sex revealed interlandmark distances that vary significantly between the parental populations. Similar results were obtained using an ANOVA on linear distances calculated from the landmark coordinates in the entire study sample of 131 individuals of varying admixture proportion. Regression analyses identified correlations between the absolute distance (in mm) between the landmarks and proportion of West African ancestry after adjustment for variation due to sex and height. Many interlandmark distances showed highly significant relationships with genetic ancestry which were in some cases stronger (R2 = 0.337) than those previously identified between skin color and West African ancestry proportion (R2 = 0.211). ANOVAs were conducted using each candidate gene in combination with each interlandmark distance using sex, height and ancestry as covariates. Five out of seven candidate genes returned significant results, identifying trends of significance across regions of the face. COL11A1 shows association with normal variation around the mouth. FGFR2 appears to be associated with face length. TRPS1 displays a slight association with variation around the eyes and brow ridge. LMNA has a striking association with projection of the face. BRAF shows association with variation regarding the nose, notably with nose width. Landmarking trials conducted by independent observers returned similar results, with comparable trends. The directionality of morphological change was investigated using ƒÒƒz the regression coefficient, to characterize change in interlandmark distance corresponding with change in allelic composition. Much of these results were in concordance with data presented in previous studies that described the size ratio of traits between West African and European individuals. The results of this study show there is great potential for elucidation of the genetic basis of normal facial feature variation. The associations shown with specific facial traits and patterns of variation suggest there is strong reason to further investigate these candidate genes and other genes of interest.