Unravelling the genetic basis of stuttering: GWAS meta-analysis highlights link with rare speech disorders

Abstract

Developmental stuttering affects up to 11% of children, with around one-fifth developing a persistent stutter. Twin and family studies indicate a strong genetic contribution and comorbidity with other heritable traits, but the common genetic architecture of clinically ascertained stuttering, persistence and recovery remains poorly characterised. We performed a genome-wide association study (GWAS) meta-analysis of 6,096 cases and 81,629 controls of European ancestries from 18 cohorts, including 2,953 cases from stuttering-specific collections with clinically confirmed or validated self-reported phenotypes. We additionally analysed persistence, severity and frequency, and integrated our results with an independent GWAS of self-reported stuttering. Our meta-analysis and the self-report GWAS showed high genetic correlation (rg = 0.66). A polygenic score from the self-report GWAS was associated with stuttering in our ascertained cases, with modest predictive value ({Delta}AUC = 0.067, p = 1.8 x 10-{superscript 2}3), and with greater severity and frequency among people who stutter. We replicated two self-report loci (CAMTA1, CTNND2), and combined analysis identified a genome-wide significant association at MPPED2. Our 24 suggestive loci implicated genes with established roles in childhood apraxia of speech, including SETBP1, CAMTA1, PTBP2, GRIN2A and KIRREL3, with strong enrichment for apraxia genes (OR = 40.1, p = 1 x 10-). Within our meta-analysis, persistent stuttering was genetically near-identical to stuttering overall (rg = 0.97). Genetically predicted brain imaging nominally implicated basal ganglia and cerebellar motor circuitry. Common-variant liability to stuttering is shared across ascertainment approaches, relates to severity, and converges with rare-variant genes for speech disorders on circuits underlying speech motor control.