High-resolution oligonucleotide array mapping of 15q12-q13.3 rearrangements (chr15:25,700,000–31,400,000). Although there appears to be variation in the exact location of breakpoints, all map to large blocks of segmental duplication at BP3, BP4, and BP5 (indicated by dashed lines). For each individual, deviations of probe log₂ ratios from zero are depicted by grey/black lines, with those exceeding a threshold of 1.5 standard deviations from the mean probe ratio coloured green and red to represent relative gains and losses, respectively. Segmental duplications of increasing similarity (90–98%, 98–99%, and >99%) are represented by grey/yellow/orange bars, respectively. A number of other 15q rearrangements with breakpoints mapping to BP3, BP4, and BP5 are shown in Supplementary Figure 5.

Pedigrees and patient photographs of 15q13 deletions. Developmental delay and seizure phenotype is indicated by left- and right-half shaded symbols, respectively. Presence or absence of 15q13 deletions is shown below each symbol in all individuals tested (absence of text indicates unavailable for testing). Photographs of affected family members are below each pedigree. We obtained consent to publish photographs from each individual included in this figure. (a) Family of proband IMR338. All affected individuals have 3.9 Mb deletions. Note the full everted lips and deep-set eyes evident in affected individuals. IMR338Cb is unaffected and does not have the deletion. (b) Patient 02961 (de novo deletion); note hypertelorism, synophrys, prominent philtrum, everted upper lip, and hypotonic facies. (c) Patient 69/06 (de novo deletion); note the prominent philtrum, everted upper lip, hypertelorism, and hypotonic facies. (d) Family of proband CMS5826. Note upslanting palpebral fissures and prominent philtrum in the patient.

Duplication architecture of 15q13 breakpoint regions. (a) Paralogy between large (≥10 kb), highly identical (≥95%) segmental duplications (blue bars) is shown between BP3, BP4 and BP5 as pairwise alignments (blue lines). Sequence assembly gaps are shown as purple bars. (b) The underlying duplicon structure (blocks of identical colour represent those that share the same evolutionary origin) and orientation (red/black arrows) of pairwise alignments between the blocks9. (c) RefSeq genes. BP3-BP4 and BP3-BP5 share fewer large, high-identity duplications when compared to BP4-BP5 (BP4-BP5, total aligned bp=571.8 kb, mean identity=98.6%) (Supplementary Table 1). Most notable are three segmental duplications each with >99.4% identity ranging in length from 96 kb to 218 kb, in which the breakpoints of all recurrent 1.5 Mb deletions we describe occur (red arrows). All three of these large duplications lie in an opposing orientation between BP4 and BP5 in the reference assembly. As a result, inversions of this region could result in these duplications being placed in a direct orientation, creating a configuration predisposed to microdeletion by NAHR.

Duplications of (a) 15q13.1-q13.2 (BP3-BP4), and (b) 15q13.3 (BP4-BP5) identified in 2 of 960 normal control samples using the HumanHap300 Genotyping BeadChip (Illumina, San Diego, CA). Data shows probe position (x-axis) against smoothed LogR intensity ratio (red/black bars) and B-allele frequency (blue/black circles) (y-axis) for probes on chromosome 153. Data points within the duplicated regions are shown in colour, while those outside are shown in black. Green shaded boxes indicated segmental duplications. Within the limits of resolution of this data, the BP4-BP5 duplication appears to be the reciprocal event to the recurrent deletion shown in Figure 1. However, the BP3-BP4 duplication is clearly smaller than the deletion identified in patient 543/06, with breakpoints located in unique sequence (not in segmental duplications). No deletions of BP3-BP4 or BP4-BP5 were observed in this control population.

Identification of a common inversion polymorphism in 15q13.3. To detect inversions of the BP4-BP5 microdeletion region, we performed FISH mapping using fosmid probes located proximal and distal to BP5. The separation of these probes in the reference assembly is ~2.5 Mb, enough to visualize these as two separate signals on metaphase chromosomes. Inversion of the region BP4-BP5 moves the two probes within close proximity to each other, visualized as overlapping yellow signals. Using this assay on 8 HapMap individuals of different ethnicities, we observed the inversion on 7/16 chromosomes (~44% of the population, Supplementary Table 2). We also tested the mother of patient 69/06 (in whose germline the BP4-BP5 deletion arose), who was found to be heterozygous. Shown are individuals (a) homozygous for the reference allele, (b) heterozygous, or (c) homozygous for the inversion allele.