Childhood apraxia of speech (CAS) is a rare neurodevelopmental speech disorder affecting approximately 0.1% of the population. It's characterized by deficits in speech planning and programming, leading to inaccurate sequencing of sounds and syllables and impaired prosody, resulting in unintelligible speech. The gene *FOXP2* was the first gene identified as being involved in CAS etiology in 2001, but for nearly two decades, it was the only one implicated in cases without intellectual disability. Recent advances in genome sequencing and bioinformatics have made it possible to identify additional genes involved in CAS. Two independent cohort studies have identified variants in a total of 17 genes associated with CAS, with a combined diagnostic yield of 37%. These studies revealed that many children with CAS have a single-gene diagnosis and implicated genes involved in transcriptional regulation, suggesting a critical role for transcriptional dysregulation in speech development. Additional pathways, such as G-protein signaling pathways, were also implicated. These studies also demonstrated that pathogenic variants are more often de novo than inherited, highlighting the genetic heterogeneity of speech disorders, similar to other neurodevelopmental disorders. The identified genes were frequently associated with other neurodevelopmental disorders, such as epilepsy and intellectual disability. These findings emphasize the need for larger studies to further identify causal genes, improve diagnostic yield and understand molecular pathways involved in CAS. This study aimed to identify the molecular basis of CAS in a larger cohort of probands, performing comprehensive phenotypic analysis and genome sequencing to identify pathogenic variants. The study also investigated the molecular co-expression of genes associated with CAS and the overlap of these genes with those associated with other neurodevelopmental disorders.

Previous research has established a strong genetic component in childhood apraxia of speech (CAS). The discovery of *FOXP2* mutations as a cause of CAS was a landmark finding, but it only explained a fraction of cases. Two significant studies, Eising et al. (2019) and Hildebrand et al. (2020), utilized genome-wide sequencing to identify additional genes. Eising et al. identified eight genes (CHD3, SETD1A, WDR5, KAT6A, SETBP1, ZFHX4, TNRC6B, and MKL2) in 19 probands (42% diagnostic rate), while Hildebrand et al. implicated nine additional genes (CDK13, EBF3, GNAO1, GNB1, DDX3X, MEIS2, POGZ, UPF2, and ZNF142) in 33 probands (33% diagnostic rate). These studies revealed a shared pathway involvement in transcriptional regulation, particularly genes like POGZ, SETBP1, SETD1A, and KAT6A, which highlighted transcriptional dysregulation's importance in abnormal speech development. Other relevant pathways identified included G-protein signaling pathways (GNAO1, GNB1). The studies showed de novo mutations were more common than inherited ones, further highlighting genetic heterogeneity. Many identified genes had associations with other neurodevelopmental disorders, such as epilepsy and intellectual disability. This background underscores the need for expanded research to improve understanding of CAS's genetic basis and to find more precise therapies.

This study involved 70 unrelated probands with a primary diagnosis of CAS, excluding those with moderate to severe intellectual disability. Participants were recruited via clinicians or direct parental referral. Comprehensive phenotypic analysis was conducted, including medical and developmental history, and assessments of speech, language, and cognition using standardized tools. Diagnosis of CAS was confirmed using the American Speech-Language-Hearing Association's consensus criteria. Genomic DNA was extracted from blood or saliva samples. Trio genome sequencing was performed on 204 individuals from the 70 families (71 probands, 127 parents, and 6 other relatives). Illumina sequencing platforms were used, aiming for 30-fold depth. Variant analysis incorporated trio or parent-child pair designs. Read mapping, quality control, and variant calling were performed using standard bioinformatics pipelines. Variant filtering focused on rare variants (gnomAD allele count ≤2), absence in unaffected family members, and consistent inheritance models. Variants were annotated using the Variant Effect Predictor (VEP). Loss-of-function (LoF) and predicted damaging missense variants were analyzed. A two-stage approach was used for shortlisting candidate variants, prioritizing variants in genes of interest (previously implicated genes and genes associated with relevant neurodevelopmental disorders) followed by a genome-wide search. ACMG guidelines were used for pathogenicity assessment, supplemented by clinical geneticist review. High-confidence variants were classified as pathogenic or likely pathogenic, while low-confidence variants were classified as uncertain significance or likely pathogenic but with inconsistencies in phenotype or evidence of pathogenicity. Copy number and structural variants were also analyzed using Manta and qDNAseq. High-confidence variants were validated with Sanger sequencing or ddPCR. Additional analyses included short tandem repeat (STR) analysis, polygenic risk score (PRS) analysis for ASD and non-syndromic cleft palate, and mitochondrial gene abundance estimation. Finally, brain gene co-expression analysis was conducted using data from the BrainSpan Atlas, and gene set enrichment analysis was performed to identify enriched pathways.

High-confidence variants were identified in 18 of 70 (26%) probands. These variants were found in 18 different genes, including 15 novel genes (ARHGEF9, BRPF1, DIP2C, ERF, HNRNPK, KDM5C, PHF21A, RBFOX3, SETD1B, SHANK3, SPAST, TAOK2, TRIP12, and ZBTB18), three genes already linked to CAS (SETD1A, DDX3X, SETBP1), The variants included frameshift, splice acceptor site, nonsense, missense variants and an exon duplication. Fifteen of these variants were de novo, and three were inherited. Many of the newly implicated genes, along with previously identified CAS-related genes, are associated with other neurodevelopmental disorders (ASD, epilepsy, intellectual disability). Phenotypic analysis revealed that probands with high-confidence variants more frequently exhibited other neurodevelopmental features compared to those without variants. Three forms of novel genetic analysis were undertaken: STR analysis, polygenic risk score (PRS) analysis for ASD and non-syndromic cleft palate and estimation of mitochondrial gene abundance. No significant findings were identified using the polygenic risk score, and the mitochondrial analysis did not suggest that abundance is a biomarker for CAS. Co-expression analysis using BrainSpan data revealed significant co-expression between the newly identified genes during brain development, suggesting a common biological pathway. Gene set enrichment analysis demonstrated that genes involved in chromatin organization and transcriptional regulation were significantly over-represented among the CAS-associated genes. The study prioritized additional candidate genes beyond the high-confidence set by re-examining low-confidence variants from this study and previous work, and then by analysing 21 large copy number variant regions from the literature. This analysis identified several genes associated with chromatin organization and/or DNA binding which could be causal for CAS.

This study significantly expands the number of genes associated with CAS, providing evidence for the involvement of 15 novel genes and confirming the role of three previously identified genes. The high diagnostic yield (26%) underscores the significant contribution of monogenic causes to CAS. The enrichment of genes involved in chromatin organization and transcriptional regulation points to critical biological mechanisms underlying speech development. The substantial overlap between genes implicated in CAS and other neurodevelopmental disorders suggests shared genetic etiologies and potential pleiotropic effects. The observation that probands with genetic diagnoses more frequently exhibited additional neurodevelopmental impairments suggests a possible threshold effect. However, this requires further investigation. Although many genes identified were previously linked to neurodevelopmental disorders, the study expands the phenotypic spectrum of those genes, specifically linking them to CAS, highlighting the importance of detailed phenotypic characterization in genetic studies. The co-expression analysis supports a model where alterations in chromatin remodeling and transcriptional regulation disrupt the developmental processes crucial for speech.

This study nearly doubles the number of known genes associated with CAS, highlighting the significant genetic contribution to this disorder. The identification of 15 novel genes and the confirmation of three previously implicated genes strongly supports the involvement of chromatin organization and transcriptional regulation in speech development. The high diagnostic yield is comparable to other neurodevelopmental disorders, emphasizing the importance of genetic testing in CAS. Future research should focus on larger cohorts to validate findings, investigate functional consequences of the identified variants, and explore the potential for precision medicine approaches.

The study's strict criteria for defining high-confidence variants might have led to an underestimation of the true number of causal genes. The filtering strategy might have been overly conservative for recessive inheritance models. The sample size, while larger than previous studies, may not be sufficient to fully capture the genetic heterogeneity of CAS. Further investigation is needed to fully understand the functional consequences of the identified variants and their precise roles in speech development.