SURVEY AND SUMMARY Computational disease gene identification: a concert of methods prioritizes type 2 diabetes and obesity candidate genes

Genome-wide experimental methods to identify disease genes, such as linkage analysis and association studies, generate increasingly large candidate gene sets for which comprehensive empirical analysis is impractical. Computational methods employ data from a variety of sources to identify the most likely candidate disease genes from these gene sets. Here, we review seven independent computational disease gene prioritization methods, and then apply them in concert to the analysis of 9556 positional candidate genes for type 2 diabetes (T2D) and the related trait obesity. We generate and analyse a list of nine primary candidate genes for T2D genes and five for obesity. Two genes, LPL and BCKDHA, are common to these two sets. We also present a set of secondary candidates for T2D (94 genes) and for obesity (116 genes) with 58 genes in common to both diseases. INTRODUCTION Genome-wide empirical studies generate large sets of potential candidate genes. However, it remains difficult to identify the most likely disease-related genes. In this study, we review existing computational methods for disease gene identification and describe their differences and similarities, and then we apply these methods in a complementary fashion to the identification of candidate disease genes for the complex disease type 2 diabetes (T2D) and the related trait obesity. Our aim is to offer the prospective user an overview of the inputs, outputs and functionality of the methods, to illustrate the use of these methods applied to a single problem, and to present a new set of most likely candidate disease genes for the complex diseases T2D and obesity. Familial studies indicate that many diseases are caused by ectopic loss or gain of gene function, and research to date has identified numerous genes implicated in simple (Mendelian) diseases. Existing methods have been successful in identifying single high relative risk disease genes, but have typically failed to identify genes underlying complex diseases or traits *To whom correspondence should be addressed. Tel: +27 21 9592611; Fax: 27 21 9592512; Email: [email protected]