An estimated 22 million children worldwide under age five are overweight. In these children, obesity is a primary indicator for development of type 2 diabetes and possibly cancer. I present a research program, SELECTionPREDISPOSED, to identify novel obesity-risk genes as tools for detection of early childhood obesity making possible a selective prevention program in predisposed children. I will use records and blood samples from children and their parents in the Mother-Child Cohort of Norway, the Health Survey of Nord-Trøndelag and the Norwegian Birth Registry and cross-correlate the databases for genetic research. I hypothesize that children large at birth with enhanced infantile growth may be predisposed to obesity by genetic factors. Obesity-linked genes are likely to include a mix of variants associated with glucose, insulin and fat metabolism and may be identifiable in population studies using biobanks and end-point registries. The state-of-the-art approach is to identify diabetes- or obesity-associated genes in subjects with disease. My approach is to investigate subsets of children with high and low birth weights and BMIs at age six. Using cutting-edge genetic techniques like GWAS, copy-number variation and massive parallel exome and epigenome sequencing I will correlate the genetic information with clinical data in large national end-point registries by a case-control design subsequent replication in large data sets and control for environmental confounders by cross-correlation to the national birth registry. I want to change the field by working with predisposed children in order to influence the ratio between those that may and may not develop obesity and diabetes. In this way my team will develop contextual tools of a groundbreaking nature. This “tool-kit” may make it possible to identify and implement in predisposed children, an early low-cost prevention program to slow down and reverse the development of obesity and prevent diabetes and possibly cancer.