Background Molecular diagnosis of neurodevelopmental disorders (NDDs) is mainly based on exome sequencing (ES), with a diagnostic yield of 31% for isolated and 53% for syndromic NDD. As sequencing costs decrease, genome sequencing (GS) is gradually replacing ES for genome-wide molecular testing. As many variants detected by GS only are in deep intronic or non-coding regions, the interpretation of their impact may be difficult. Here, we showed that integrating RNA-Seq into the GS workflow can enhance the analysis of the molecular causes of NDD, especially structural variants (SVs), by providing valuable complementary information such as aberrant splicing, aberrant expression and monoallelic expression. Methods We performed trio-GS on a cohort of 33 individuals with NDD for whom ES was inconclusive. RNA-Seq on skin fibroblasts was then performed in nine individuals for whom GS was inconclusive and optical genome mapping (OGM) was performed in two individuals with an SV of unknown significance. Results We identified pathogenic or likely pathogenic variants in 16 individuals (48%) and six variants of uncertain significance. RNA-Seq contributed to the interpretation in three individuals, and OGM helped to characterise two SVs. Conclusion Our study confirmed that GS significantly improves the diagnostic performance of NDDs. However, most variants detectable by GS alone are structural or located in non-coding regions, which can pose challenges for interpretation. Integration of RNA-Seq data overcame this limitation by confirming the impact of variants at the transcriptional or regulatory level. This result paves the way for new routinely applicable diagnostic protocols.
/Summary: Genes encoding long non-coding RNAs (lncRNAs) comprise a large fraction of the human genome, yet haploinsufficiency of a lncRNA has not been shown to cause a Mendelian disease. CHASERR is a highly conserved human lncRNA adjacent to CHD2– a coding gene in which de novo loss-of-function variants cause developmental and epileptic encephalopathy. Here we report three unrelated individuals each harboring an ultra-rare heterozygous de novo deletion in the CHASERR locus. We report similarities in severe developmental delay, facial dysmorphisms, and cerebral dysmyelination in these individuals, distinguishing them from the phenotypic spectrum of CHD2 haploinsufficiency. We demonstrate reduced CHASERR mRNA expression and corresponding increased CHD2 mRNA and protein in whole blood and patient-derived cell lines–specifically increased expression of the CHD2 allele in cis with the CHASERR deletion, as predicted from a prior mouse model of Chaserr haploinsufficiency. We show for the first time that de novo structural variants facilitated by Alu-mediated non-allelic homologous recombination led to deletion of a non-coding element (the lncRNA CHASERR ) to cause a rare syndromic neurodevelopmental disorder. We also demonstrate that CHD2 has bidirectional dosage sensitivity in human disease. This work highlights the need to carefully evaluate other lncRNAs, particularly those upstream of genes associated with Mendelian disorders.
Le diagnostic moléculaire des troubles neurodéveloppementaux (TND) repose encore principalement sur le séquençage d’exome, avec un rendement diagnostique plafonnant à 31 % pour les TND isolés et 53 % pour les TND syndromiques. Le séquençage de génome (SG) a montré de meilleures performances mais sa mise en œuvre systématique en diagnostic pose de nouveaux défis, notamment en ce qui concerne la détection et l'interprétation des variants structuraux (SV) ou des variants localisés dans des régions non codantes.Nous avons montré que l'application de techniques omiques complémentaires, telles que le RNA-Seq clinique et la cartographie optique du génome, contribuent à améliorer le rendement diagnostique du séquençage du génome pour des cas complexes ou difficiles à interpréter. Notre étude confirme que, sur la base du SG seul, certains SV n’auraient pas été retenus comme pathogènes/candidats et qu’il aurait été difficile de poser les hypothèses mécanistiques pour les événements complexes. Par ailleurs, nous avons identifié et rapporté une délétion du long ARN non codant régulateur CHASERR que nous associons à un nouveau TND syndromique.Le RNA-Seq clinique et les techniques de séquençage long fragments sont des sujets de recherche en pleine expansion. Leur application en diagnostic de routine n’est pas encore standardisée et nécessitera encore une longue mise au point. Avec nos résultats nous apportons un éclairage nouveau sur cette introduction progressive, en association avec le SG.