henhaobio
Sign in

NatureGenetic & Genomic MedicinePeer-reviewed journal article

De novo variants in the non-coding spliceosomal snRNA gene RNU4-2 are a frequent cause of syndromic neurodevelopmental disorders

Yuyang CheniD, Ruebena DawesiD, Hyung Chul KimiD, Alicia LjungdahliD, Sarah L. StentoniD, Susan WalkeriD, Jenny LordiD, Gabrielle LemireiD, Alexandra C Martin-GearyiD, Vijay S GaneshiD, Jialan MaiD, Jamie M. EllingfordiD, Erwan DelageiD, Elston N. D’SouzaiD, Shan DongiD, David R. AdamsiD, Kirsten AllaniD, Madhura Bakshi, Erin E. BaldwiniD, Seth BergeriD, Jonathan Adam BernsteiniD, Ishita Bhatnagar, Ed Blair, Natasha J. BrowniD, Lindsay Catherine BurrageiD, , David H. ViskochiliD

PDFDOIPreprintData
In-depth full-text digest available · Jump to digest ↓

Abstract

Around 60% of individuals with neurodevelopmental disorders (NDD) remain undiagnosed after comprehensive genetic testing, primarily of protein-coding genes1. Increasingly, large genome-sequenced cohorts are improving our ability to discover new diagnoses in the non-coding genome. Here, we identify the non-coding RNA RNU4-2 as a novel syndromic NDD gene. RNU4-2 encodes the U4 small nuclear RNA (snRNA), which is a critical component of the U4/U6.U5 tri-snRNP complex of the major spliceosome2. We identify an 18 bp region of RNU4-2 mapping to two structural elements in the U4/U6 snRNA duplex (the T-loop and Stem III) that is severely depleted of variation in the general population, but in which we identify heterozygous variants in 119 individuals with NDD. The vast majority of individuals (77.3%) have the same highly recurrent single base-pair insertion (n.64_65insT). We estimate that variants in this region explain 0.41% of individuals with NDD. We demonstrate that RNU4-2 is highly expressed in the developing human brain, in contrast to its contiguous counterpart RNU4-1 and other U4 homologs, supporting RNU4-2s role as the primary U4 transcript in the brain. Overall, this work underscores the importance of non-coding genes in rare disorders. It will provide a diagnosis to thousands of individuals with NDD worldwide and pave the way for the development of effective treatments for these individuals.

In-depth summary

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (9 Apr 2024) · CC BY 4.0

Mostly de novo variants in an 18 bp constrained region of U4 snRNA gene RNU4-2 cause a syndromic neurodevelopmental disorder, estimated to explain 0.41% of cases.

About 60% of individuals with neurodevelopmental disorders (NDD) stay undiagnosed after genetic testing centred on protein-coding genes, and non-coding RNA genes have been little examined. Using large-scale genome sequencing, the work adds a dominant disorder of a major-spliceosome snRNA to the known recessive minor-spliceosome snRNA disorders (RNU12, RNU4ATAC), in a region that commercial clinical exomes miss.

Key results

  1. ShownA single-base insertion, n.64_65insT, recurs de novo in undiagnosed NDD probands more often than the most recurrent protein-coding variant in a large developmental disorder cohort. It is essentially absent from population and non-NDD cohorts.

    46 GEL carriers (38 de novo in trios; eight others consistent with de novo) = 0.52% of 8,841 undiagnosed NDD probands; absent in 3,408 diagnosed NDD, 21,817 non-NDD and 33,122 unaffected GEL individuals, 76,215 gnomADv4.0 and 245,400 All of Us genomes; one UK Biobank carrier (AF 1.02×10^-6). PACS1 p.Arg203Trp, top coding recurrence, is in 36/31,058 DD (0.12%). · A highly recurrent insertion explains 0.52% of…; Multiple factors likely explain the high…

  2. ShownAn 18 bp single-stranded U4 segment (nucleotides 62 to 79, spanning the T-loop and the stem III site that pairs with U6) is depleted of population variation but carries variants in 119 individuals with NDD, 92 (77.3%) of them with n.64_65insT.

    UK Biobank shows 26% of possible SNVs in the region vs a 78% median across the rest of RNU4-2. Single-base insertions: 54/8,841 (0.61%) undiagnosed GEL NDD probands vs 2/490,132 UK Biobank (OR=1,531; P=3.3×10^-92). Any variant: 60/8,841 (0.68%) vs 39/490,132 (OR=85.8; 95%CI:56.4-131.6). Remainder of gene: OR=0.96, P=0.67. · The n.64_65insT variant is within a highly…; Assessing variant depletion

  3. ShownCarriers share a syndromic phenotype: moderate-to-severe developmental delay and intellectual disability, late or ataxic walking, little or no speech, dysmorphic facies, microcephaly, short stature, seizures, hypotonia, white matter abnormalities on MRI and involvement of several organ systems.

    Versus 12,203 other GEL NDD probands, n.64_65insT carriers are enriched for microcephaly (n=26; OR=6.62; P=7.87×10^-10), drooling (n=7; OR=19.2; P=2.83×10^-7), hypotonia (n=18; OR=3.60) and global developmental delay (n=37; OR=3.56). In 36 deeply phenotyped individuals, 82% walked, at a mean of 3.6 years; ≥75% had short stature, microcephaly, seizures and hypotonia. · A highly recurrent insertion explains 0.52% of…; Variants in this crucial region cause a severe…

  4. SuggestedRNU4-2, not its adjacent 97.2% identical paralog RNU4-1, appears to be the main U4 transcript in developing cortex. This offers a candidate explanation for why NDD variants concentrate in RNU4-2.

    BrainVar bulk RNA-seq (176 prefrontal cortex samples, 6 post-conception weeks to 20 years): expression of the two genes is tightly correlated but consistently higher for RNU4-2. ATAC-seq of two prenatal cortex samples (18 and 19 gestational weeks) shows far stronger accessibility at RNU4-2. The constrained centre of RNU4-1 shows no NDD enrichment in GEL. · RNU4-2 is highly expressed across tissues and in…; Analysing RNU4-2 and RNU4-1 expression; Prenatal prefrontal cortex ATAC-seq data

  5. ShownFollow-up analyses found no transcriptome-wide splicing defect in carriers' blood. They also found no other NDD-associated snRNA genes or constrained snRNA subregions among brain-expressed snRNAs.

    Blood RNA-seq of five carriers (three n.64_65insT, one n.77_78insT, one n.78A>C) vs 5,409 controls: no excess OUTRIDER expression outliers or FRASER2 retained-intron/outlier events. Of 12 non-RNU4-2 snRNAs tested for de novo burden, all Fisher's P>0.15; among constrained regions only RNU4-2 passed (P=1.34×10^-9; 33 vs 0 probands), others P>0.25. · RNA-sequencing in blood does not show a global…; No other spliceosomal snRNA genes are enriched…; RNA-sequencing of individuals with RNU4-2…; Burden testing and statistical analysis

Methods, data and reproducibility

Case-control and de novo burden testing in Genomics England genomes (undiagnosed NDD vs non-NDD probands and UK Biobank), an 18 bp sliding-window SNV-depletion scan of 28 brain-expressed snRNA genes, multi-cohort carrier collection with deep phenotyping, blood RNA-seq outlier calling, and brain expression and chromatin accessibility profiling.

  • Genomics England v18 (100,000 Genomes Project) participants grouped by HPO/ICD-10 codes; 10,388 undiagnosed NDD individuals, 8,841 of them probands · human · n = 13,812 NDD; 21,817 non-NDD probands; 33,122 unaffected
  • GEL high-confidence de novo callset (stringent_filter) and aggregated variant call set aggV2 · human · n = 13,949 trios (12,554 probands); aggV2 29,850 probands
  • Population genome datasets used as controls and for depletion scoring: UK Biobank, gnomAD v4.0, All of Us · human · n = UKB 490,640; gnomAD 76,215; All of Us 245,400
  • Additional NDD carriers from NHS GMS, GREGoR, Undiagnosed Diseases Network, Broad CMG, Matchmaker Exchange and collaborators; detailed clinical data for 36 of all 119 carriers · human · n = NHS GMS 21; GREGoR 10; UDN 8; personal/MME 17
  • Autism trios from SSC, SPARK and MSSNG searched for critical-region variants (3/7,149 carriers) · human · n = 7,149 ASD trios; 4,180 sibling control trios
  • GEL whole-blood RNA-seq (globin/rRNA depleted, 100 bp paired-end) of RNU4-2 carriers vs controls · human · n = five carriers; 5,409 controls
  • BrainVar bulk RNA-seq of postmortem dorsolateral prefrontal cortex or frontal cerebral wall, 6 post-conception weeks to 20 years · human · n = 176 samples (167 DLPFC, 9 frontal wall)
  • ATAC-seq of prenatal telencephalic wall/prefrontal cortex at 18 and 19 gestational weeks · human · n = two samples
Methods & tools
Fisher's exact test (R) with Bonferroni correction, 18 bp sliding-window SNV depletion metric (UK Biobank), Monte Carlo Fisher-Pitman permutation test, OUTRIDER (expression outliers), FRASER2 (splicing outliers) via DROP pipeline, Illumina DRAGEN alignment, STAR aligner and DEXSeq gene counts (GENCODE v31), ENCODE ATAC-seq pipeline, bedtools subtractBed, IGV and UCSC Genome Browser (variant and mappability checks), Ensembl v111 snRNA annotation, RCSB PDB cryo-EM structure 6QW6
Reagents & environment
Only data/annotation versions stated: GEL v18, gnomAD v4.0, Ensembl v111, GENCODE v31, GRCh38.p12 (STAR), T2T CHM13v2.0/hs1 mappability check, NHS GMS V3; no software versions or containers given.
Reproducibility gaps
Data statement says all data are in the manuscript; individual-level GEL, UK Biobank and clinical cohort genomes are not released, and no analysis code or repository is provided.
Limitations (6)
  • Splicing was assessed only in blood RNA from five carriers. Brain tissue was unavailable, and paralogs such as RNU4-1 might compensate in blood, so tissue-specific or subtle splice-site effects remain untested. · stated by the authors
  • The search for other snRNA disease genes had low power because the genes and constrained regions are short (mean 139.5 bp and 28.1 bp), and recessive inheritance was not examined. · stated by the authors
  • The 0.41% population estimate assumes a 40% upstream diagnostic rate. The ARID1B comparison cohort lacks some long-established syndromic genes, so comparisons of relative frequency are approximate. · stated by the authors
  • Why n.64_65insT recurs so often is unresolved. Local mutability, secondary structure and germline selection are hypotheses only, and testing selection would require deep sequencing of testes or sperm. · stated by the authors
  • No functional experiments (e.g. U4/U6 unwinding, spliceosome assembly or cellular models) test the proposed structural mechanism; disrupted positioning of U6 ACAGAGA is inferred from structure alone. · noted in this summary
  • Additional carriers were gathered largely via personal communication, and the 36 deeply phenotyped individuals are mostly n.64_65insT carriers (33), so the phenotypes of rarer insertions and SNVs are weakly characterised. · noted in this summary

Read this if you analyse genomes from patients with undiagnosed neurodevelopmental disorders, study spliceosomal snRNA biology, or want a constraint approach for short, highly mutable non-coding genes.

Citations
156
FWCI
41.22
Field percentile
top 1%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISRius '25Nava '25Porras '25Greene '26Leitão '26Jonghe '25
83% semantic matchGenetic & Genomic Medicine
Rocío Rius, Alexander J. M. Blakes, Yuyang Chen+71 more·medRxiv·Aug 13, 2025· 5 citations