Classification and genetic analysis of human facial features

This thesis considers normal person to person variation in facial features, as opposed to aberrant variation usually classified as dysmorphology. An attempt has been made to categorise the face simply, as a combination of a relatively small number of variable components, in an attempt to reduce the complexity of individual appearance and make the face more amenable to genetic analysis. Three-dimensional facial data have been collected from over 1000 individuals, both unrelated individuals and family groups, using two types of optical surface scanner and the merits of these systems were evaluated. Methods of analysing the data were investigated, a traditional landmark approach was used for direct measurement of various facial dimensions and novel techniques exploring properties of surface geometry were used to delineate the face into areas of similar shape to facilitate discrete classification. Aspects of midline features including the chin cleft and nose were classified into discrete categories and analysed at both the population and family level. When tested for Mendelian segregation in available family data (two-parent plus offspring families), the nose band classification fitted to an autosomal dominant model of inheritance (40 families) and the chin cleft fitted to an autosomal codominant model (49 families studied). The majority of subjects analysed were normal individuals but a small set of patients (n=14) with Medullary Sponge Kidney (MSK) were investigated in order to explore face scanning as an objective diagnostic tool. Minor asymmetry of the face, previously described in MSK, was assessed using two landmark-based approaches but was not found to be significant compared with matched controls. The analysis and measurements made to establish phenotypic classification of the face have enabled preliminary genotype-phenotype association studies to be performed. Polymorphisms in 2 candidate genes were typed and analysed in a set of up to 97 scanned unrelated subjects where DNA had been collected. Although no significant associations were found this demonstrated the practical feasibility of such an approach for identifying genes responsible for facial variation. This project provides a novel perspective on addressing the complexity of the genetics of facial variation. The classifications described in this thesis have made aspects of individuality amenable to association and family studies and open the way for the development of larger-scale initiatives to identify genes responsible for facial appearance