Obesity: from animal models to human genetics to practical applications.

Although many animal models are used in genetic studies, the mouse is most common. Analysis of single-gene mutations, linkage analysis in crossbred strains, and gene targeting are the primary techniques used to associate obesity phenotypes with specific genes or alleles. The orthologous human gene can then be tested, either in linkage studies in families or in genome-wide association studies (GWAS), for effect on the phenotype. Frequent lack of concordance between mouse and human obesity genes may be due to the difference in phenotypes measured in humans (body mass index) versus mouse (fat mass or % body fat), lack of intermediate phenotypes, and the fact that identified genes account for only a small percentage of the heritability of common obesity, suggesting that many genes remain unknown. New technology allows analysis of individual genomes at a reasonable cost, making large-scale obesity genome projects in humans feasible. Such projects could identify common allelic variants that contribute to obesity and to variable individual response to obesity therapy. Currently, family history may be more predictive than genetics for risk of obesity, but individual testing could ultimately guide therapy and, in the aggregate, guide public health policy. The primary limitation to development of genotype-based diets is that successful randomized diet trials of widely ranging macronutrient content, adequately powered for finding rare Mendelian mutations, have not been performed.