Genetics of Eating Disorders
: a candidate gene and a genome-wide association approach

Student thesis: Doctoral ThesisDoctor of Philosophy

Abstract

Eating behaviour is an important aspect of social behaviour, and illnesses like eating disorders thus have a profound impact on quality of life. Family and twin studies provide compelling evidence for the heritability of eating disorders; it is estimated that roughly half of the phenotypic variance is accounted for by genetic factors. The presumed genetic architecture of eating disorders constitutes multiple genetic variants, each with a small effect size. This thesis aimed to replicate findings from the most extensive genetic studies of eating disorders done thus far, in a sample of 700 anorexia nervosa cases and 700 controls from the United Kingdom, the Netherlands, and Austria. The results are non-significant, which is in line with genetic studies of other psychiatric disorders, as well as complex traits such as human intelligence, height, and body mass index. Power to detect genetic risk variants with small effect sizes could be increased by larger samples sizes, and by focussing on disease related quantitative traits rather than diagnoses. Candidate quantitative traits for eating disorders include drive for thinness, bulimia, and body dissatisfaction as measured by the Eating Disorders Inventory (EDI) questionnaire. Chapter 3 of this thesis presents the distribution of these traits in a general population sample from the United Kingdom (n = 3,624 females), and Chapter 4 presents the results of genome-wide association gene (GWAG) analyses of these traits. No gene p values passed a multiple gene testing correction, but among the top genes were several previously implicated in the aetiology of eating disorders. Larger sample sizes would be needed to verify these results. The results of this thesis underscore the phenotypic and aetiologic complexity of eating disorders, but demonstrate that a general population approach using quantitative trait measurements combined with genome-wide hypothesis-free gene analyses can be fruitful.
Date of Award1 Jul 2012
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorDavid Collier (Supervisor) & Ulrike Schmidt (Supervisor)

Cite this

'