Genome-wide analyses of ADHD identify 27 risk loci, refine the genetic architecture and implicate several cognitive domains

Abstract

Attention deficit hyperactivity disorder (ADHD) is a prevalent childhood psychiatric disorder, with a major genetic component. Here we present a GWAS meta-analysis of ADHD comprising 38,691 individuals with ADHD and 186,843 controls. We identified 27 genome-wide significant loci, which is more than twice the number previously reported. Fine-mapping risk loci highlighted 76 potential risk genes enriched in genes expressed in brain, particularly the frontal cortex, and in early brain development. Overall, ADHD was associated with several brain specific neuronal sub-types and especially midbrain dopaminergic neurons. In a subsample of 17,896 exome-sequenced individuals, we identified increased load of rare protein-truncating variants in cases for a set of risk genes enriched with likely causal common variants, suggesting implication of SORCS3 in ADHD by both common and rare variants. We found ADHD to be highly polygenic, with around seven thousand variants explaining 90% of the SNP heritability. Bivariate gaussian mixture modeling estimated that more than 90% of ADHD influencing variants are shared with other psychiatric disorders (autism, schizophrenia and depression) and phenotypes (e.g. educational attainment) when both concordant and discordant variants are considered. Additionally, we demonstrated that common variant ADHD risk was associated with impaired complex cognition such as verbal reasoning and a range of executive functions including attention.

Working memory and reaction time variability mediate the relationship between polygenic risk and ADHD traits in a general population sample

“…21,22,25 Our results show that such relationships are also evident in a populationbased sample of children, linking poorer cognitive functioning in each of the three phenotypes with increased ADHD traits. The association between higher ADHD PRSs and poorer working memory accuracy as well as increased reaction time variability are consistent with prior findings where ADHD-PRS was significantly associated with decreased working memory in an independent cohort, 11 and other studies linking dysfunctions of these mechanisms to ADHD genetic risk factors, 3,82,83 providing further weight to the conceptualisation of working memory and reaction time variability as endophenotypes for ADHD. Partial mediation observed for working memory confirms previous results from a case-control study, 42 but contrasts the full mediation identified for working memory linking ADHD PRSs and symptoms of hyperactivity/impulsivity, 14 suggesting that gene-cognition-trait relationships might differ between ADHD symptom domains.…”

“…Participants of European ancestry were selected for all further analyses in order to match the genetic ancestry of the discovery genome wide association study (GWAS) for ADHD used to calculate PRSs. 11, 47 To maximise the sample size for each analysis, three partially overlapping samples at 2-year follow-up were selected to ensure consistency of age across samples (details are provided in the following sections; see Figure 1 for participant demographics). Consistent with the overall prevalence of ADHD in the general population, 48 5.3% of all subjects with available parent-rated Child Behaviour Checklist (CBCL) 49 attention problems scores measured at the 2-year follow-up ( n = 5,823) would be considered to have clinical ADHD (based on 65 t-score cut-off).…”

“…In that study the ADHD polygenic risk load (ADHD-PRS) was significantly associated with several measures of cognition, such as degreased attention, working memory and verbal reasoning in 8,722 individuals from the Philadelphia Neurodevelopment Cohort. 11 Individual differences in executive functions have been found to remain stable across development despite some overall group-level improvements from childhood through to adolescence, 40 qualifying them as candidate trait-like endophenotypes. Collectively, these results suggest that working memory, response inhibition and reaction time variability may serve as endophenotypes for the dimensional study of ADHD traits.…”

Linking GWAS to pharmacological treatments for psychiatric disorders

“…For each disorder GWAS [ADHD (8), schizophrenia (6), major depression (10), bipolar disorder (54), diabetes (57), heart failure (58), rheumatoid arthritis (41), and inflammatory bowl disease (59)], we constructed a score vector, , across 2155 genes that quantified the similarity of each gene to variants identified in each GWAS. The similarity was defined based on four main bioinformatic datasets including SNP position on the DNA, protein interaction network (PPI), brain eQTL, and regional gene expression across the brain as quantified using Allen Human Brain Atlas (AHBA) (47).…”

Genome-wide analysis of binge-eating disorder identifies the first three risk loci and implicates iron metabolism

“…Following these steps, 387,392 samples with European ancestry and 557,369 variants were retained. Processing and imputation performed on the PNC genotypes has also been previously characterized, and resulted in 4,973 European ancestry samples and 4,903,082 variants retained 64 . The ABCD Data Analysis, Informatics & Resource Center (DAIRC) performed initial genotype QC according to the Ricopili pipeline recommendations, resulting in 11,099 samples and 516,598 variants retained in data release 3.0 65 .…”