A Palindrome-Like Structure on 16p13.3 Is Associated with the Formation of Complex Structural Variations and SRRM2 Haploinsufficiency

Abstract

SRRM2 encodes a splicing factor recently implicated in developmental disorders due to a statistical enrichment of de novo mutations. Using data from the 100,000 Genomes Project, four unrelated individuals with intellectual disability (ID) were identified, each harbouring de novo whole gene deletions of SRRM2. Deletions ranged between 248 and 482 kb in size and all distal breakpoints clustered within a complex 144 kb palindrome situated 75 kb upstream of SRRM2. Strikingly, three of the deletions were complex, with inverted internal segments of 45-94 kb. In one proband-mother duo, de novo status was inferred by haplotype analysis. Together with two additional patients who harboured smaller predicted protein-truncating variants (p.Arg632^∗ and p.Ala2223Leufs^∗13), we estimate the prevalence of this condition in cohorts of patients with unexplained ID to be ~1/1300. Phenotypic blending, present for two cases with additional pathogenic variants in CASR/PKD1 and SLC17A5, hampered the phenotypic delineation of this recently described condition. Our data highlights the benefits of genome sequencing for resolving structural complexity and inferring de novo status. The genomic architecture of 16p13.3 may give rise to relatively high rates of complex rearrangements, adding to the list of loci associated with recurrent genomic disorders.

Figure 1

Complex structural rearrangements and genomic architecture at the SRRM2 locus. (a) IGV read alignments and subway plots showing 4 de novo structural rearrangements leading to SRRM2 haploinsufficiency. Region shown is 16 : 2,418,234-3,200,000 (GRCh38). Horizontal blue and green arrows indicate noninverted and inverted chromosome segments, respectively, whilst dotted red lines indicate junctions. For Family 3, a second possible configuration that could explain the short-read data is indicated. (b) Dot plot using reference sequence for the same region shown in (a) highlights a palindrome of ~144 kb (red arrow/box). A zoomed-in dot plot is shown in Figure S1.

Figure 2

Inference of de novo status of deletion by haplotype analysis. Regions shown are 16 : 3071987-3072027 16 : 3076654-3076693 16 : 3080121-3080161 16 : 3089289-3089328 16 : 3089345-3090809 (GRCh38). Read alignments show 6 SNPs inside the deleted region which are hemizygous for the alternate allele in the proband but homozygous for the reference allele in the mother. The nondeleted chromosome in the proband must have been inherited from the father. The panel on the rights highlights the proximal deletion breakpoint where split read-pairs (blue) and a drop in coverage are seen only in data from the proband. In contrast with the exonic NEDD4L deletion (Figure S2a) there was no indication of any mosaicism.

Figure 3

Dysmorphic features seen in P3 (a–e) and P5 (f–j). Clinical photographs showing (a) P3 aged 16 y with geographic tongue, (b) small chin and large ear, (c) thin upper lip and bulbous nose, (d) small hands, and (e) small feet with large halluces. (f) P5 aged 3 y and 3 m with smooth philtrum, thin upper lip, and full cheeks. (g) P5 aged 9 y and 9 m showing smooth philtrum, full cheeks, (h) fleshy ear lobes, (i) tapering fingers, and (j) long halluces compared to other toes.