Discovery of genetic factors for reading ability and dyslexia

Abstract

The ability to read is critical to access wider learning and achieve qualifications, for accessing employment, and for adult life skills. Approximately one in ten individuals are affected by dyslexia, a learning difficulty which primarily impacts word reading and spelling. Specifically, phonological processing (the ability to decode phonemes) is impaired in dyslexia. Whilst some believe dyslexia represents the extreme end of a continuum of reading ability, others have suggested it is a distinct trait.

Variation in reading ability is a highly heritable (possibly 70%) complex trait caused by many genetic variants with a small effect size. However, the genetic architecture of reading ability and dyslexia is largely unknown due to a lack of quantitative genetic studies with sufficient statistical power to detect such small effect sizes. Previously, most genetic studies of reading ability have been conducted using samples of children with dyslexia, which tend to be modest in size. Whilst large samples of genotyped unselected adults have been collected (for example UK Biobank), phenotypic data on reading or language skills is rarely prioritised.

The overall aim of this thesis is to discover genetic variants associated with dyslexia and variation in reading skill in order to better understand the aetiology of reading difficulties, which in turn, may inform prediction, identification and intervention strategies in the future. Firstly, I will conduct a genome-wide association (GWA) study of over 50,000 adults with a self-reported dyslexia diagnosis and over 1 million controls to identify associated single nucleotide polymorphisms (SNPs). I will also explore ways to improve power for discovering genetic factors associated with reading ability. To do this, I will first investigate whether unselected adult samples are valid as a means to identify genetic factors associated with reading skill through a candidate gene approach. Secondly, I will investigate whether proxy reading phenotypes are also a means to gain power through large cohorts that have no quantitative measure of reading ability. Such samples may be informative for future GWA meta-analysis of quantitative reading ability.

In Chapter 1, I will first introduce reading ability and dyslexia. I will discuss how reading ability is a quantitative trait and how it can be measured before discussing how dyslexia is identified. Then, I will consider how dyslexia may relate to reading ability: whether it represents the extreme end of a continuum of reading or whether it is a distinct trait. I will then introduce the known causes of variation in reading ability and dyslexia, which includes both environmental and genetic factors. Next, I will present the history of genetic studies of reading ability and dyslexia and their limitations. Finally, I will discuss the current state of genetic research into reading ability and introduce the aims of my thesis in detail.

Chapter 2 is a publication in Nature Genetics entitled ‘Discovery of 42 genome-wide significant loci associated with dyslexia’ which includes GWA analysis of over 1 million 23andMe, Inc participants reporting on dyslexia diagnosis. I identify 42 independent genome-wide significant loci, 15 of which are in genes previously linked to cognitive ability and/or educational attainment, and 27 of which are novel and may be more specific to dyslexia. Extensive downstream biological analysis is performed alongside genetic correlations with other traits and dyslexia polygenic score prediction of quantitative reading scores.

Chapter 3 is a publication in Twin Research and Human Genetics on ‘The association of dyslexia and developmental speech and language disorder candidate genes with reading and language abilities in adults’ which analyses an adult population cohort with quantitative measures of reading and language ability to replicate previous associations of candidate genes and biological pathways with dyslexia. I demonstrate that unselected adult populations are a valid means by which to identify genes which have previously been associated with dyslexia and/or speech and language disorder.

Chapter 4 is a research chapter in which I construct a proxy reading phenotype from measures of reading frequency in an unselected adult sample for whom a quantitative measure of reading ability is not available. I find that a dyslexia polygenic score constructed from the dyslexia GWA analysis in Chapter 3 cannot explain variation in the proxy phenotype suggesting that book reading is not a sufficient substitute for reading ability.

Finally, in Chapter 5, I integrate and discuss my research findings. I highlight the discovery of 42 variants associated with dyslexia through GWAS, in addition to the discovery of new genes and biological pathways which may form part of the biological basis of dyslexia. Following this, I consider what GWAS tells us about candidate gene findings. I discuss traits which are genetically correlated with dyslexia, including quantitative reading skills and ADHD. I consider the relationship between dyslexia and reading ability, and how genetic studies can help us to understand this better. I also consider the relationship between dyslexia and other developmental disorders, and how genetic studies can help us to understand this better. Lastly, I discuss methods to boost power for GWAS of reading ability.