Genome sequencing provides high diagnostic yield and new etiological insights for intellectual disability and developmental delay

Abstract

Short-read genome sequencing (GS) is a powerful technique for investigating the genetic etiologies of rare diseases, capturing diverse genetic variations that are challenging to approach with exome sequencing (ES). We performed GS on 260 families with intellectual disability/developmental delay. GS detected potentially disease-related variants in 55 of the 260 families, with structural resolution by long-read sequencing or optical genome mapping, and functional assessment by RNA sequencing. Excluding 31 theoretically ES-resolvable cases, GS yielded likely pathogenic variants in 17 of 229 as well as variants of unknown significance in 7 of 229, totaling 10.5%. These variants implicated several new etiological mechanisms: a microduplication syndrome involving ATP6V0C; disturbed interactions of TBL1XR1 and NR2F1 with putative cis-regulatory elements by chromosomal rearrangements; and a CCG repeat expansion near the CHD3 transcription start site. This study highlights the critical role of GS in clinical diagnostics and its potential to advance understanding of genetic disorders.

Fig. 1. Overview of this study.

This figure was created in BioRender. Hamanaka, K. (2025) https://BioRender.com/f32v286.

Fig. 2. De novo variants in non-coding genes RNU4-2 and CHASERR.

a De novo small variants in the critical region of RNU4-2 in Pt0712 and Pt2092. Lollipop plot: Pathogenic/Likely pathogenic variants in ClinVar (https://ftp.ncbi.nlm.nih.gov/pub/clinvar/vcf_GRCh38/clinvar_20250608.vcf.gz). Green rectangle: RNA secondary structure. Heatmap plot: aggregated allele counts of single-nucleotide variants at each genomic position, derived from gnomAD v4.1.0 (https://gnomad.broadinstitute.org/data). b De novo deletion affecting CHASERR in Pt2190. Shown is an IGV view of read coverage and reads supporting the deletion. Read coloring follows the conventions of IGV for paired-end alignments (https://igv.org/doc/desktop/#UserGuide/tracks/alignments/paired_end_alignments).

Fig. 3. De novo high-copy-number amplification of ATP6V0C.

A de novo 16p13.3 multiplication in Pt2074. a Upper, copy number profiles estimated using CNView. Lower, adjoined breakends identified through GS are illustrated as curved lines in the IGV snapshots. SegDup is displayed as in the UCSC genome browser (https://genome.ucsc.edu/cgi-bin/hgTrackUi?g=genomicSuperDups). The copy number profiles are presented in bins indicated on the top right corner, with gray intervals representing the mean ± two standard deviations of 1330 samples. Read colors follow IGV conventions. b The overall structure estimated from optical genome mapping data. Breakpoints, inferred from optical genome mapping data, are colored black. c Bar plot of gene expression levels in RNA-seq of Pt2074’s LCL (replicate n = 2) relative to control samples (n = 17). Individual replicates and control samples are indicated by black dots over the red and blue bars, respectively. d Two DECIPHER cases with similar 16p13.3 duplications to Pt2074. Blue bar: duplication; black bar: multiplication in Pt2074. c, d Gene models are taken from the gnomAD constraint metrics track of the UCSC genome browser (https://genome.ucsc.edu/cgi-bin/hgTrackUi?db=hg19&g=gnomadPLI) and color-coded according to LOEUF.

Fig. 4. De novo intergenic SVs disturbing interactions of TBL1XR1 and NR2F1 with putative cis-regulatory elements.

Chromatin landscape around TBL1XR1 (a) and NR2F1 (c). From top to bottom, the figure displays the contact map and chromatin loop calls of an HFFc6 micro-C, coverage tracks of ATAC-seq and histone ChIP-seq including H3K27ac, H3K4me3, and H3K4me1 modifications in BE(2)-C and SH-SY5Y, and GENCODE V44 gene models colored according to UCSC conventions (https://genome.ucsc.edu/cgi-bin/hgTrackUi?hgsid=2280748172_eutv8PURT0qZPmMVtttVc16EEnGI&c=chr12&g=wgEncodeGencodeV44). Dotted lines indicate SV breakpoints observed in Pt2286 (a) and Pt2800 as well as in a previous study (c). b, d Enlarged views of the (a) and (c) panels. The top tracks depict the regions interacting with the TBL1XR1 TSS (b) or NR2F1 TSS (d). To the left of the coverage tracks are the depth-normalized coverage intervals on the Y-axis.

Fig. 5. De novo DLX4 retrotransposition possibly altering splicing in STXBP1 deep intronic region.

A de novo retrotransposition of a DLX4 transcript into STXBP1 intron 10 in Pt2178. Depicted are IGV snapshots of STXBP1 (top) and DLX4 (middle) alongside GENCODE transcript models, and SpliceAI score distribution along the mutant STXBP1 intron 10 sequence in Pt2178 (bottom). For the SpliceAI score, the upper portion illustrates donor site scores while the lower portion displays acceptor site scores. Blue bars: wild-type sequence scores; red bars: mutant sequence scores; blue rectangle: potential pseudo-exonization based on SpliceAI predictions.

Fig. 6. De novo intronic CCG repeat expansion near the CHD3 TSS.

a Counts of in-CCG-repeat reads anchored to Chr17:7884501-7886140 in CHD3 among Pt481, his parents, and healthy control individuals. b–d T-LRS of the Chr17:7,885,308-7,885,345 CCG repeat at CHD3 in Pt481 and his mother. b Frequency of CCG repeat size in T-LRS reads. Red, forward read; blue, reverse read. c Waterfall plot showing the tri- or di-nucleotide composition of the CCG repeat expansion and flanking sequences in T-LRS reads. Tri- or di-nucleotides were colored as shown in the upper right corner. d DNA methylation status around the predominant CHD3 TSS. Top, GENCODE transcript models; middle, total CAGE counts aggregated from various cells and tissues (https://genome.ucsc.edu/cgi-bin/hgTrackUi?db=hg38&g=fantom5); bottom, T-LRS reads with (“exp.”) or without (“no-exp.”) CCG repeat expansion. Horizontal line: T-LRS read; black circle: methylated CpG; white circle: unmethylated CpG.