Obesity is a highly heritable disorder but the genetic associations reported to date account for only a small percentage of the inherited variation in body mass index. Two groups report deletions on chromosome16p11.2 that may explain part of the 'missing heritability' in terms of 'high-penetrance' mutations that are rare but when present are very often associated with severe obesity. This is in contrast to more common gene defects that are less closely associated with clinical symptoms. Bochukova et al. identified rare recurrent copy number variants in 300 patients with severe early-onset obesity, caused by deletions involving several genes including SH2B1, known to be involved in leptin and insulin signalling. Many of the patients also suffered neurodevelopmental disorders. Walters et al. identified deletions of at least 593 kilobases on chromosome 16p11.2 in 31 patients with a previously unrecognized type of extreme obesity. The strategy they used to identify the lesion — using small well-phenotyped cohorts of extreme phenotypes with targeted follow-up in genome-wide association studies and population cohorts — shows promise as a means of identifying 'missing heritability' in complex metabolic diseases more generally. Recently, numerous single nucleotide polymorphisms have been identified as being associated with obesity, but these loci together account for only a small fraction of the known heritable component. Here, an association is reported between rare deletions of at least 593 kilobases at 16p11.2 and a highly penetrant form of obesity. The strategy used of combining study of extreme phenotypes with targeted follow-up is promising for identifying missing heritability in obesity. Obesity has become a major worldwide challenge to public health, owing to an interaction between the Western ‘obesogenic’ environment and a strong genetic contribution1. Recent extensive genome-wide association studies (GWASs) have identified numerous single nucleotide polymorphisms associated with obesity, but these loci together account for only a small fraction of the known heritable component1. Thus, the ‘common disease, common variant’ hypothesis is increasingly coming under challenge2. Here we report a highly penetrant form of obesity, initially observed in 31 subjects who were heterozygous for deletions of at least 593 kilobases at 16p11.2 and whose ascertainment included cognitive deficits. Nineteen similar deletions were identified from GWAS data in 16,053 individuals from eight European cohorts. These deletions were absent from healthy non-obese controls and accounted for 0.7% of our morbid obesity cases (body mass index (BMI) ≥ 40 kg m-2 or BMI standard deviation score ≥ 4; P = 6.4 × 10-8, odds ratio 43.0), demonstrating the potential importance in common disease of rare variants with strong effects. This highlights a promising strategy for identifying missing heritability in obesity and other complex traits: cohorts with extreme phenotypes are likely to be enriched for rare variants, thereby improving power for their discovery. Subsequent analysis of the loci so identified may well reveal additional rare variants that further contribute to the missing heritability, as recently reported for SIM1 (ref. 3). The most productive approach may therefore be to combine the ‘power of the extreme’4 in small, well-phenotyped cohorts, with targeted follow-up in case-control and population cohorts.
