Several limitations of this study should be taken into account when interpreting its results. It is likely that there is a selection bias in the sample of patients studied. Both in France and in Germany, patients were recruited at OCD outpatient clinics, where individuals with severe developmental disabilities and associated medical conditions are unlikely to come, thus decreasing the possibility of detecting pathogenic CNVs [30]. Second, because the purpose of our study was to detect the large deletions that are typically observed in Prader-Willi and DiGeorge syndromes, we did not screen for small deletions in the 15q11-q13 and 22q11 regions (none of which has been shown to be pathogenic) or for intragenic deletions. An additional limitation of our study is the relatively limited size of our sample. As pathogenic CNVs are rare events, type II errors could explain our inability to detect any rearrangement in the 15q11-q13 and 22q11 regions.
