Candidate gene analyses are often inconclusive owing to genetic or phenotypic heterogeneity, low statistical power, selection of nonfunctional SNPs, and inadequate statistical analysis of the genetic architecture. Angiotensin-converting enzyme (ACE) is involved in adipocyte growth and function and the ACE-processed angiotensin II inhibits adipocyte differentiation. Associations between body mass index (BMI) and ACE polymorphisms have been reported in general populations, but the contribution to severe obesity of this gene, which is located under an obesity genome-scan linkage peak on 17q23, is unknown. ACE is one of the most studied genes and markers responsible for variation in circulating ACE enzyme levels have been extensively characterised. Eight of these variants were genotyped in 1054 severely obese cases and 918 nonobese controls, as well as 116 nuclear families from the genome scan (n=447), enabling the known clades to be inferred. Qualitative analysis of individual single-nucleotide polymorphisms (SNPs), haplotypes, clades, and diploclades demonstrated no significant associations (P<0.05) after minimal correction for multiple testing. Quantitative analysis of clades and diploclades for BMI, waist-to-hip ratio, or ZBMI in children were also not significant. This rigorous, large-scale study of common, well-defined, severe polygenic obesity provides strong evidence that functionally relevant sequence variation in ACE, whether it is defined at the level of SNPs, haplotypes, or clades, is not associated with severe obesity in French Caucasians. Such a study design exemplifies the strategy needed to clearly define the contribution of the ACE gene to the plethora of complex genetic diseases where weak associations have been previously reported.