Background: The obesity epidemic is the greatest public health problem of our time, and exerts an enormous health and economic burden by acting as a risk factor for multiple disorders and all-cause mortality. While environmental and social factors certainly contribute to the complex etiology of obesity, there is strong evidence of a substantial genetic component. The majority of obesity genes are involved the leptin-melanocortin receptor pathway governing energy homeostasis, but mutations affecting this circuit are often untreatable and rare, and an improved understanding of other genetic risk factors could aid in the development of novel therapies. In this thesis I study two obesity candidate genes with unclear direct relevance to disease: 1) rare structural variation at the 16p11.2 BP4-BP5 locus and 2) common variation in the Fat Mass and Obesity-Associated (FTO) gene. Methods: 1) I analyzed disinhibited eating measurements from families with 16p11.2 copy number variation (CNV) carriers, to test whether eating in the absence of hunger (EAH) and loss of control (LOC) eating behaviors mediate the dosage-dependent CNV-BMI relationship. 2) Using association data from a study of over 20,000 African Americans and 1,145 functional annotations from the Encyclopedia of Non-coding Elements (ENCODE) and Roadmap Epigenomics projects, I statistically fine-mapped the FTO locus to identify the SNP(s) and cellular contexts underlying the association between FTO and obesity.

Results: 1) EAH due to external triggers mediates over 30% of the 16p11.2 deletion’s effect on obesity, while other EAH and LOC behaviors were not significant mediators. This result was independent of IQ deficits and autism related to the CNV, as well as parents’ feeding behaviors and practices. 2) Given 51 FTO SNPs’ association statistics, correlation, and overlap with functional annotations, rs9927317 and rs62033405 had the highest posterior probability of association with obesity. Obesity-associated SNPs may regulate expression of FTO and/or nearby genes through the activity of enhancers and 5’ ends of transcribed genes in the substantia nigra of the brain, bone chondrocytes, and white adipose. Conclusions: These results may help pinpoint the specific genes, regulatory elements, and cellular contexts through which the 16p11.2 and FTO loci exert their effects on obesity.