Recently, the application of array-based comparative genomic hybridization (array CGH) has improved rates of detection of chromosomal imbalances in individuals with mental retardation and dysmorphic features. Here, we describe three individuals with learning disability and a heterozygous deletion at chromosome 17q21.3, detected in each case by array CGH. FISH analysis demonstrated that the deletions occurred as de novo events in each individual and were between 500 kb and 650 kb in size. A recently described 900-kb inversion that suppresses recombination between ancestral H1 and H2 haplotypes encompasses the deletion. We show that, in each trio, the parent of origin of the deleted chromosome 17 carries at least one H2 chromosome. This region of 17q21.3 shows complex genomic architecture with well-described low-copy repeats (LCRs). The orientation of LCRs flanking the deleted segment in inversion heterozygotes is likely to facilitate the generation of this microdeletion by means of non-allelic homologous recombination.

Prader-Willi syndrome (PWS) can result from a 15q11-q13 paternal deletion, maternal uniparental disomy (UPD), or imprinting mutations. We describe here the phenotypic variability detected in 51 patients with different types of deletions and 24 patients with UPD. Although no statistically significant differences could be demonstrated between the two main types of PWS deletion patients, it was observed that type I (BP1-BP3) patients acquired speech later than type II (BP2-BP3) patients. Comparing the clinical pictures of our patients with UPD with those with deletions, we found that UPD children presented with lower birth length and started walking earlier and deletion patients presented with a much higher incidence of seizures than UPD patients. In addition, the mean maternal age in the UPD group was higher than in the deletion group. No statistically significant differences could be demonstrated between the deletion and the UPD group with respect to any of the major features of PWS. In conclusion, our study did not detect significant phenotypic differences among type I and type II PWS deletion patients, but it did demonstrate that seizures were six times more common in patients with a deletion than in those with UPD.

Angelman syndrome (AS) can result from either a 15q11 -q13 deletion (del), paternal uniparental disomy (UPD), imprinting, or UBE3A mutations. Here, we describe the phenotypic and behavioral variability detected in 49 patients with different classes of deletions and nine patients with UPD. Diagnosis was made by methylation pattern analysis of exon 1 of the SNRPN-SNURF gene and by microsatellite profiling of loci within and outside the 15q11 -q13 region. There were no major phenotypic differences between the two main classes (BP1 -BP3; BP2 -BP3) of AS deletion patients, except for the absence of vocalization, more prevalent in patients with BP1 -BP3 deletions, and for the age of sitting without support, which was lower in patients with BP2 -BP3 deletions. Our data suggest that gene deletions (NIPA1, NIPA2, CYF1P1, GCP5) mapped to the region between breakpoints BP1 and BP2 may be involved in the severity of speech impairment, since all BP1 -BP3 deletion patients showed complete absence of vocalization, while 38.1% of the BP2 -BP3 deletion patients were able to pronounce syllabic sounds, with doubtful meaning. Compared to UPD patients, deletion patients presented a higher incidence of swallowing disorders (73.9% del  22.2% UPD) and hypotonia (73.3% del  28.57% UPD). In addition, children with UPD showed better physical growth, fewer or no seizures, a lower incidence of microcephaly, less ataxia and higher cognitive skills. As a consequence of their milder or less typical phenotype, AS may remain undiagnosed, leading to an overall underdiagnosis of the disease.

We report array-CGH screening of 95 syndromic patients with normal G-banded karyotypes and at least one of the following features: mental retardation, heart defects, deafness, obesity, craniofacial dysmorphisms or urogenital tract malformations. Chromosome imbalances not previously detected in normal controls were found in 30 patients (31%) and at least 16 of them (17%) seem to be causally related to the abnormal phenotypes. Eight of the causative imbalances had not been described previously and pointed to new chromosome regions and candidate genes for specific phenotypes, including a connective tissue disease locus on 2p16.3, another for obesity on 7q22.1→q22.3, and a candidate gene for the 3q29 deletion syndrome manifestations. The other causative alterations had already been associated with well-defined phenotypes including Sotos syndrome, and the 1p36 and 22q11.21 microdeletion syndromes. However, the clinical features of these latter patients were either not typical or specific enough to allow diagnosis before detection of chromosome imbalances. For instance, three patients with overlapping deletions in 22q11.21 were ascertained through entirely different clinical features, i.e., heart defect, utero-vaginal aplasia, and mental retardation associated with psychotic disease. Our results demonstrate that ascertainment through whole-genome screening of syndromic patients by array-CGH leads not only to the description of new syndromes, but also to the recognition of a broader spectrum of features for already described syndromes. Furthermore, on the technical side, we have significantly reduced the amount of reagents used and costs involved in the array-CGH protocol, without evident reduction in efficiency, bringing the method more within reach of centers with limited budgets.

No abstract