Several lines of evidence suggest that the aetio-pathogenesis of the common form of type 2 diabetes mellitus and its intrinsically related features of impaired insulin secretion and decreased insulin sensitivity (insulin resistance) includes a strong genetic component. At present, however, little is known about the nature of this genetic component although familial clustering of the disease has been described for decades. Major break-throughs in the genetic sciences of type 2 diabetes have been identifications of insulin receptor gene mutations in syndromes of severe insulin resistance and mutations in pancreatic beta-cell genes in the monogenic sub-group of type 2 diabetes: maturity-onset-diabetes-of-the-young, MODY. Pathophysiological models of insulin resistance in skeletal muscles and impaired glucose-induced insulin secretion in the beta-cells have formed a basis for selecting candidate genes with potential influence on the development of type 2 diabetes ("diabetogenes"). This process of selecting and analyzing genes for mutations that potentially associate with either type 2 diabetes mellitus, insulin resistance or impaired insulin secretion is often described as the "candidate gene approach". The studies reported in this thesis are excerpts from an extensive strategy of genetically dissecting (mutation analysis) in: 1) patients with the common form of late-onset type 2 diabetes mellitus the pathways that transduce the insulin signals from the plasma membrane to the activation of glycogen synthesis in skeletal muscle, and in 2) patients with either late-onset type diabetes or MODY the pathways involved in normal beta-cell development and beta-cell function (insulin secretion). Twelve of the genes that encode proteins in the insulin-signalling pathway from the insulin receptor through the phosphatidylinositide-regulated kinases down to the complex of phosphatases that regulate glycogen synthesis in skeletal muscle were analyzed. We could not confirm that a Val985Met variant in the insulin receptor is associated with type 2 diabetes or that the Met326Val of the p85 alpha regulatory subunit of the phosphoinositide-3 kinase is associated with insulin resistance. We found no coding mutations (missense) in the insulin signalling protein kinases but we confirmed that the 5 bp deletion (PP1ARE) in the 3'-end of the PPP1R3 gene that encodes the glycogen-associated regulatory subunit of protein phosphatase-1 (PP1G) is associated with insulin resistance estimated as insulin mediated glucose uptake. In contrast to protein kinases in skeletal muscles the genes encoding beta-cell transcription factors (IPF-1, NeuroD1/BETA2, and Neurogenin 3) are polymorphic but we could not confirm that the Asp76Asn of IPF-1 is a susceptibility gene for late-onset type 2 diabetes. On the other hand we confirmed that the Ala45Thr variant in NeuroD1/BETA2 may represent a susceptibility gene for type 1 diabetes but none of these genes revealed any MODY-specific mutations. Also the gene encoding the ATP-regulatable potassium channels of the beta-cell (Kir6.2) is polymorphic but none of these polymorphisms associated with changes in glucose-induced insulin secretion. Reviewed in context of the existing data our studies support the candidate gene approach as a feasible method for directly either identifying or excluding any gene as a diabetes-susceptibility gene ("diabetogene").