Genetic variants and phenotypic data curated for the CAGI6 intellectual disability panel challenge

Abstract

Neurodevelopmental disorders (NDDs) are common conditions including clinically diverse and genetically heterogeneous diseases, such as intellectual disability, autism spectrum disorders, and epilepsy. The intricate genetic underpinnings of NDDs pose a formidable challenge, given their multifaceted genetic architecture and heterogeneous clinical presentations. This work delves into the intricate interplay between genetic variants and phenotypic manifestations in neurodevelopmental disorders, presenting a dataset curated for the Critical Assessment of Genome Interpretation (CAGI6) ID Panel Challenge. The CAGI6 competition serves as a platform for evaluating the efficacy of computational methods in predicting phenotypic outcomes from genetic data. In this study, a targeted gene panel sequencing has been used to investigate the genetic causes of NDDs in a cohort of 415 paediatric patients. We identified 60 pathogenic and 49 likely pathogenic variants in 102 individuals that accounted for 25% of NDD cases in the cohort. The most mutated genes were ANKRD11, MECP2, ARID1B, ASH1L, CHD8, KDM5C, MED12 and PTCHD1 The majority of pathogenic variants were de novo, with some inherited from mildly affected parents. Loss-of-function variants were the most common type of pathogenic variant. In silico analysis tools were used to assess the potential impact of variants on splicing and structural/functional effects of missense variants. The study highlights the challenges in variant interpretation especially in cases with atypical phenotypic manifestations. Overall, this study provides valuable insights into the genetic causes of NDDs and emphasises the importance of understanding the underlying genetic factors for accurate diagnosis, and intervention development in neurodevelopmental conditions.

Fig. 1

Number of pathogenic/ likely pathogenic (Blu bars) and Variants of Uncertain Significance (green bars) identified in the cohort of 415 individuals. The ANKRD11 gene exhibits the highest number of pathogenic/likely pathogenic variants, while several other genes show a mix of both variant categories

Fig. 2

Mosaic variant in the GRIN2A gene. A IGV visualization of the mosaic missense variant detected in GRIN2A, analyzed using the Ion Torrent PGM platform on DNA extracted from a blood sample (top) and oral mucosa cells (bottom). B Structure of the human GluN1/GluN2A NMDA receptor in the glutamate/glycine-bound state (left) and GluN2A NMDA receptor molecule (right) (PDB code: 6IRA). The Arginine 504 residue is indicated as a purple sphere

Fig. 3

De novo intronic variant c.4956-17A in MED13L. A Localisation of GRIN2A variant in the genomic sequence and its predicted effect on splicing process; B Transcript analysis of the intronic variant in MED13L performed for the control sample (top) and patient’s sample (bottom). The intronic variant generates a new Acceptor splicing site leading to the inclusion of 16 intronic nucleotides (intron 21) into exon 22; C Pedigree showing the absence of intronic variant in healthy parents; D The predicted premature truncated protein, p.Ser1652Argfs*1, will miss the MID domain of the Med PIWI module, the core globular domain of MED13 protein