Obesity is a heritable complex trait (1,2). Genome-wide association studies (GWASs) led to the identification of various common single-nucleotide polymorphisms (SNPs) for human obesity (3–8). Although most of the results of GWASs have been highly reproducible, they explain only a minor fraction of the variance of body mass index (BMI) when compared with the total expected heritability of BMI [∼50%, (9)]. Hence, a substantial ‘missing heritability' (10) becomes obvious, which might in part be explained by copy number variants (CNVs). CNVs are by definition chromosomal regions with sizes of 1kb to several megabases (Mb) being interindividually present in variable numbers. At a genome-wide level, thousands of CNVs have already been identified. Owing to resolution of the current technology, most of these CNVs are >5 kb (11). The database of Genomic Variants (http://projects.tcag.ca/variation/) currently lists 57 829 CNVs at 14 478 CNV loci and half of these CNVs are of sizes from 1 to 10 kb. The frequency spectrum and the precise structure of CNVs are closely related to the technical and algorithmic methods applied (12,13). Specific CNVs have been related
