GWAS studies have been successful in finding genetic determinants of obesity. To translate discovered genetic variants into new therapies or prevention strategies, molecular or physiological mechanisms need to be discovered. One strategy is to perform data mining of data sets with detailed phenotypic data, such as those present in dbGAP (database of Genotypes and Phenotypes) for hypothesis generation. We propose a novel technique that combines the power and computational efficiency of existing Bayesian Network (BN) learning algorithms with the statistical rigor of Structural Equation Modeling (SEM) to produce an overall system that searches the space of potential networks and evaluates promising candidates using standard SEM model selection criteria. We illustrate our method using the analysis of a candidate SNP data set from the AMERICO sample, a multi-ethnic cross-sectional cohort of roughly three hundred children with detailed obesity-related phenotypes. We demonstrate our approach by showing genetic mechanisms for three obesity-related SNPs.