SU78 - GENOME--WIDE ASSOCIATION STUDY OF LEARNING DISORDERS IN THE DANISH POPULATION

Abstract

Background Learning disabilities are around 5–10 percent prevalent among children in the school-age. These disorders grouped under the term ‘Specific Developmental Disorders of Scholastic Skills’ (ICD-10: F81) demonstrate substantial aggregation within families of affected members. They co-occur frequently with other psychiatric disorders such as Attention Deficit Hyperactivity Disorder (ADHD), Autism Spectrum Disorder (ASD) and Major Depressive Disorder (MDD). Twin based studies report heritability-estimates ranging from 30 to 70 percent. Despite comprehensive multidisciplinary studies over the past decades, the specific causal mechanisms are unknown. Methods This study was based on 1,206 cases diagnosed with any learning disorder according to the ICD-10 criteria (ICD-10: F81) and 13,955 controls. The cases originate from the iPSYCH consortium established to study six major psychiatric disorders. The cases are therefore enriched for comorbidity with ADHD, ASD and MDD. Hence, we included individuals in the control group, with these disorders (but no diagnosis of learning disability) in the same proportions as observed among the cases. The diagnosis information was obtained from the Danish Central Psychiatric Register. Subsequently Guthrie cards with dried blood spots from the Danish Newborn Screening Biobank were retrieved. From the blood spots, the genomic DNA was isolated, whole genome amplified in triplicates and genotyped using the Illumina Psych chip for 500,000 markers spanning entire genome. The non-genotyped variants were imputed with 1000 genome phase-3 as reference. After extensive quality control, a GWAS was conducted for ~9.7 million variants. Only Europeans and unrelated individuals were included. Significant principal components were added as covariates to adjust for population stratification. Results We found two genome-wide significant loci; one located in chromosome 11 in the intron of the gene DUSP8 (p=4.12×10e-8) and the second in chromosome 10 in a intergenic region (p=5.12×10e-8). These two loci have not been previously demonstrated for association with reading disorders or any other trait. Functional annotation of variants located within 1MB region of the loci did not pinpoint any specific risk-gene in relation to learning disabilities. Gene based analysis using MAGMA identified no exome-wide significant gene. Transcriptome wide association approach based on imputed gene expressions using PrediXcan revealed no significant genes either. Heritability estimate based on GCTA REML analysis was 0.06 (0.04) under an assumption of five percent prevalence. Polygenic risk score analysis of 27 traits revealed a significant inverse correlation with human intelligence (p=0.00003; R2=0.002). Discussion We conducted a GWAS of reading disabilities based on the medical register data and found two genome-wide significant loci. We also show significant inverse genetic correlation between reading disabilities and human intelligence. Since the cases and controls comprise predominantly individuals with major psychiatric disorders such ADHD, ASD and MDD, our findings could reflect learning disabilities originating based on a genetic background enriched for risk-variants for psychiatric disorders and we therefore need replication before relating the findings to the general population.