Genetic abnormalities are the most common identifiable cause of unexplained MR [
11], but conventional karyotyping is unable to detect imbalances smaller than about 3-5 Mb [
12]. Smaller chromosomal abnormalities can be identified with fluorescent
in situ hybridization (FISH) or multiplex ligation-dependent probe amplification (MLPA) techniques, confirming a clinical suspicion of well-known microdeletion/microduplication syndromes (i.e. Williams syndrome, Velocardiofacial/DiGeorge syndrome, etc.) or analysing subtelomeric regions of all chromosomes. The combined analysis of karyotype and subtelomeric regions, using FISH or other molecular techniques, have allowed the detection of chromosome abnormalities in about 5-10% of these patients [
11,
12]. The newer chromosome microarray or comparative genomic hybridization technique (array-CGH) is an efficient manner to approach a case of MR. It does not require an expert clinician to suspect a specific diagnosis, and may cover the entire genome or targets known pathologic loci in an unique test, identifying deletions and/or duplications with a higher degree of sensitivity. This new technique has revealed submicroscopic chromosome aberrations in MR patients with normal results from prior cytogenetic analyses with detection rates as 5-20% [
6,
13]. However, array-CGH is incapable of detecting balanced rearrangements of chromosomal material (including reciprocal translocations and inversions), which are expected to occur in about 0.75% of all MR patients [
14]. Moreover, the interpretation of microarray data in MR is complicated by the discovery of areas of DNA segment longer than 1 kb, with a variable copy number compared with a normal reference genome, called Copy Number Variations (CNVs) [
15]. CNVs are associated to a pathological phenotype when one or more dosage-sensitive genes inside the rearranged region are altered. However, CNVs are not considered pathologic in all cases, because they appear to be conserved across primate species and may be responsible of individual diversity and human evolution. In a single individual, it is possible to detect even > 1000 non-pathological common CNVs [
16], needing the comparison with unaffected control cohorts and parental tests [
17]. The detection of a relatively large, rare,
de novo CNV in an affected patient is strongly indicative of pathological significance, and is present in about 10% of cases of MR with normal chromosome analysis [
1,
18]. CNVs should be considered causative of the condition when: 1) CNVs overlap with regions known to cause well delineated MR syndromes; 2) CNVs include the critical region of a syndrome or causative genes; 3) the phenotype of the patient is consistent with the syndrome's features [
2].