Our overall goal is to understand the factors affecting susceptibility to cardiovascular and metabolic disorders. Common forms of these disorders involve many genetic and environmental contributions, greatly complicating both genetic and biochemical approaches. To simplify the analyses, we study traits relevant to these disorders, such as atherosclerotic lesions, heart remodeling, obesity, and insulin resistance, among inbred strains of mice.
The power of mouse genetics. Over the last 100 years, a number of tools have been developed for studies with mice, such as the creation of inbred strains, the culture of embryonic stem cells and the engineering of genes. These make the mouse the most useful mammal for genetic studies. Among the hundreds of inbred strains of mice (each representing a unique gene pool in which natural variations have been fixed by inbreeding) are variations relevant to most human disorders. For example, when placed on a hyperlipidemic background, the size of atherosclerotic lesions varies more than 100-fold among the various strains. The challenge is to identify the underlying genetic differences and the pathways that they perturb.
Genetic dissection of complex traits in mice. In order to identify the genes underlying complex traits in mice, it is important to be able to map them with high precision. Traditional quantitative trait locus (QTL) analysis, involving genetic crosses between strains, has poor resolution, and has been only modestly successful. Together with the Eleazar Eskin group in Computer Sciences at UCLA, we developed an association-based strategy that has greatly improved resolution and made it possible to directly identify strong candidates (Bennett et al. , 2010). Second, to understand the pathways perturbed by the various genes, we have employed systems-based approaches along with recently developed high throughput technologies, such as expression arrays, next generation sequencing, and mass spectrometry (Schadt et al., 2005; Ghazalpour et al., 2010 ). Systems-based approaches, as contrasted with reductionistic approaches, attempt to move beyond the perspective of single genes to groups of genes. This is particularly important for studies of complex traits since they result from interactions between genes and between genes and the environment. Finally, classical gene engineering approaches can be used to test the resulting hypotheses.
Why not just study human populations? Over the last several years, human genetic studies, particularly genome-wide association studies (GWAS), have taught us a great deal about many different common disorders. But these have important limitations, such as the inability to identify interactions and to analyze molecular mechanisms. Notably, GWAS studies of most straits have succeeded in explaining a very small fraction of the genetic component. For example, GWAS of about a quarter million individuals for body mass index explained only about 1.5% of the trait variation, despite the fact that the trait has more than 50% heritability. Clearly, studies in animal models, where the genetics and environment can be controlled, will be essential for a full understanding of complex diseases.