Although GSVs at these loci were observed at similar overall frequencies to those in the original reports, we observed a low rate of replication for associations with obesity (Table 1). For 10 of the 11 GSVs detected, the reported obesity association was not replicated, even at nominal levels of significance. Of particular note was that each of the 6 GSVs originally identified in subjects with common obesity was present in at least 1 normal weight or underweight individual, contrary to the criteria used to identify these GSVs (i.e. being present exclusively in obese subjects) [16]. Indeed, for the majority of loci the GSV frequency was higher in normal weight than in obese subjects.
