Familial correlations have shown the importance of the genetic background in susceptibility to obesity (by adoption studies or studies in twins reared separately). The study of identical twins submitted to overfeeding has demonstrated the importance of the genes in response to nutritional factors, i.e. the role of the interaction genes/nutrition. The candidate genes can be selected on the basis of rodent genetic models of obesity or of metabolic studies. This approach allowed to discover the mutations involved in the very rare mendelian obesity cases, which are linked to the leptin-melanocortin pathway, Nevertheless, common obesity is a complex phenotype resulting from interactions between environmental (mainly nutritional) and genetic (polygenic) determinants. All genes whose products have effects on energy intake, energy expenditure, lipid metabolism or adipocyte biology, can be tested as candidates. Interaction can be revealed by a plethoric diet as well as a restrictive one, Some candidate genes involved in different aspects of energy or lipid metabolism modulate the susceptibility to common obesity and associated complications, and the response to low calorie diet. The Human Genome Project has provided an (almost) exhaustive list of genetic polymorphisms. These genetic variants must be tested in interaction models (longitudinal prospective studies/dietary intervention studies). These studies will help to find new physiopothologic and therapeutic ways, rather than predict on individual risk.