Approved Research
Genetics and genetic epidemiology of brain-related traits and disorders, and their interplay with physical health
Approved Research ID: 82087
Approval date: February 15th 2022
Lay summary
Mental illness is the main cause of disability in the UK and globally. Most people suffering from a mental illness do not get treated, and not everyone who receives treatment gets better. Part of the reason that some people develop mental illness while others do not is that there are differences between people in their DNA (known as genetic variants). Some genetic variants are more common in people suffering from mental illness than in people who are mentally well. These genetic variants do not cause mental illness independently, but seem to act together with features of the environment, like stress and traumatic events, to increase the risk of someone developing a mental illness. Similarly, differences in how much better people get after treatment for mental illness might also be related to genetic variants and to certain environmental features.
We will use information about mental illness and related traits in the UK Biobank to understand more about which genetic variants are more common in people with mental illness, and how they might work with environmental features to increase the risk of someone developing a mental illness. We will also study genetic variants that are more common in people who get better after treatment for mental illness compared to people who do not. We will also look at biological measures like biomarkers and metabolomics, and brain imaging, to assess their role in mental illness. We will combine our results with results from other large studies to identify these variants of interest. We will also build mathematical models to try to understand how these variants act biologically to increase the chances of developing a mental illness, or make a person more likely to get better following treatment. Finally, we will use the rich data on environmental features provided by UK Biobank to understand how the way these variants work changes when people are in different environments. Ultimately, we aim to increase our understanding of how mental illness occurs, and so provide new information for designing new treatments. We hope that what we learn from this three-year project will improve the lives of those suffering from mental illness.
Scope extension:
We will extend our previous projects (particularly #16577, #18177, #27546), conducting studies of genetic and non-genetic (environmental) factors on brain-related phenotypes, including normative behaviour, psychiatric illness, and neurological degeneration and dysfunction. Combining genotypic, biomarker, imaging, and phenotypic data, we will generate predictive models of disorder subtypes, treatment response and side effects, comorbidities, and short- and long-term disease outcomes. Related traits of interest will include socioeconomic status, traumatic experiences, and cross-domain medical comorbidities: neurological (including Alzheimer's and Parkinson's), cognitive, autoimmune, cardiometabolic, sleep, and body shape.
Specific aims include:
- Genetic associations with different definitions and subtypes of psychiatric disorders, using primary healthcare data and mental health questionnaires.
- Metabolic measures (metabolomics, laboratory parameters and biomarkers) in psychiatric disorders, particularly examining genetic control of appetite regulation, dysregulated/disordered eating behaviour, physical activity, metabolism, and food intake (extending project #27546).
- Examining common and rare structural genetic variants of varying length in psychiatric disorders.
- Predicting drug response and side effects in primary healthcare data.
- Predicting the effectiveness and complications of talking therapies (psychotherapy, counselling) and alternative approaches (mindfulness, mediation).
- Effects on onset, course, recurrence, and potential recovery of psychiatric illnesses, and their overlap with physical diseases.
- Examining genetic influences on brain structures, and their interplay with behaviours.
Extension, April 2022: We will combine the UK Biobank data with data from the National Institute for Health Research Biomedical Research Centre at the South London and Maudsley NHS Trust (NIHR BRC Maudsley), in order to carry out more powerful analyses. Specifically, we will combine UK Biobank data with:
- The Genetic Links to Anxiety and Depression study (GLAD), a nationwide study of depression and anxiety. This study has recruited >42,000 participants with >26,000 DNA samples and is much younger than UK Biobank (median age 34).
- The UK Eating Disorders Genetic Initiative (EDGI). This study is recruiting individuals with experience of an eating disorder.
- The COVID-19 Psychiatric and Neurological Genetics (COPING) study which recruited individuals during the pandemic from GLAD and EDGI as well as individuals from the NIHR BRC Maudsley without mental illness to act as controls.
- Psychiatrically well control subjects from the NIHR BRC Maudsley.
All of the NIHR BRC Maudsley data has been genotyped on the UK Biobank array, and has mental health phenotyping using an extended version of the UK Biobank mental health questionnaire. This makes the NIHR BRC Maudsley data uniquely harmonisable with UK Biobank.
Scope extension:
Identify changes in DNA that increase the risk for psychiatric disorders alone (specifically the internalising disorders: depression, anxiety including OCD, and related disorders), and for these disorders in the presence of co-morbid physical disorders (autoimmune disorders, including rheumatoid arthritis, and non-immune disorders, including type 2 diabetes, migraine, chronic pain, obesity and body-mass index). Disorder status will be determined from the UK Biobank adjudicated health outcomes, including data from primary care , hospitals, and self-report. We will also explore whether the variants associated with each psychiatric disorder predict the likelihood that an individual has a given physical disorder.
We will combine the UK Biobank data with data from the National Institute for Health Research Biomedical Research Centre at the South London and Maudsley NHS Trust (NIHR BRC Maudsley), in order to carry out more powerful analyses. Specifically, we will combine UK Biobank data with:
- The Genetic Links to Anxiety and Depression study (GLAD), a nationwide study of depression and anxiety. This study has recruited >42,000 participants with >26,000 DNA samples and is much younger than UK Biobank (median age 34).
- The UK Eating Disorders Genetic Initiative (EDGI). This study is recruiting individuals with experience of an eating disorder.
- The COVID-19 Psychiatric and Neurological Genetics (COPING) study which recruited individuals during the pandemic from GLAD and EDGI as well as individuals from the NIHR BRC Maudsley without mental illness to act as controls.
- Psychiatrically well control subjects from the NIHR BRC Maudsley.
All of the NIHR BRC Maudsley data has been genotyped on the UK Biobank array, and has mental health phenotyping using an extended version of the UK Biobank mental health questionnaire. This makes the NIHR BRC Maudsley data uniquely harmonisable with UK Biobank.
We would like to extend the scope of the research under application 16577 to examine the previously accessed phenotypes to develop methods for improving polygenic score prediction. Named delegate, Dr David Howard, will complete the research and publish a manuscript describing the work. No new bulk data categories are required, and no new novel data will be generated.
Scope extension:
GWAS methods have been developed to analyse associations between SNPs and multiple phenotypes jointly. We have produced one such method (MultiPhen) and performed a simulation study finding that multivariate analyses can double the discovery of trait associated genetic variation compared with univariate analyses. Multi-trait analyses, such as polygenic risk scores, offer insights into shared and distinct aetiology among different phenotypes, such as ADHD, autism, schizophrenia, eating disorders and obesity. We will perform single and multi-trait analyses on the UK Biobank to boost discovery power of causal genetic variants, identify shared aetiology among phenotypes, and evaluate method performance on real data.
Our expertise in multivariate methodology will enable powerful investigations of the shared aetiology of psychiatric and physical traits, including brain-related phenotypes. The release of imaging data in the UKB provides an opportunity to further integrate neuroimaging into our psychiatric research. We request the T1-MRI data to enable deeper investigation of brain-related phenotypes and their involvement in psychiatric disorders.
For example, we will generate a novel neuroimaging phenotype - 'brain-age', representing an age-adjusted index of brain health. This biomarker has been used to explore trajectories of general health in ageing as well as in psychiatric disorders. It is heritable, and therefore informative to consider as a physical trait (with psychiatric relevance) alongside other physical phenotypes in our multi-trait GWAS analysis. Thus, the shared and unique genetic risks with poorer brain health and poorer physical health can be established.
The brain-age phenotype is a single value per individual, and amenable to polygenic score analysis, to determine what genetic variation is important for brain health, and how much variability in brain-age can be explained by composite genetic measures. Brain-age is calculated using custom image analysis software that takes T1-MRI data and generates a brain-age value. These values will be returned to the UKB.
Proposed extension 25/09/18: As part of our work, we have sought to replicate previous work in Generation Scotland. To more accurately assess whether we have replicated these previous findings, we wish to remove individuals in the UK Biobank sample who are also in Generation Scotland. Accordingly, we are seeking permission to ascertain the overlap between individuals in the UKBB and the Generation Scotland datasets. We propose to apply the same procedure previously applied successfully by our collaborators as part of project 4844 (STratifying Resilience and Depression Longitudinally (STRADL), PI: Professor Andrew McIntosh). Specifically, this process converts the genotypes (sample of 50 common SNPs) of individuals in UK Biobank into a numerical identifier, the checksum. We have already received checksum data for Generation Scotland. We will match the UKB checksums against the checksums of the Generation Scotland participants. In this way, overlapping individuals can be flagged but no UKB genotype information is shared with the Generation Scotland team (not will they share their genotype information with us).
Proposed extension 06/12/22: We seek to use the UK Biobank data as a linkage disequilibrium (LD) reference panel for post-GWAS analyses, for example fine-mapping of genome-wide significant risk loci. This LD reference panel will be used for GWAS external to the UK Biobank, or multi-trait GWAS within the UK Biobank, as covered by the current project scope.
Proposed extension 2023-11-15: The existing application supports cross-phenotype genetic analyses relating to shared genetic aetiology across phenotypes. We would like to extend this application to particularly encompass social and psychiatric phenotypes. For example, we have identified significant associations of PRS for psychiatric disorders and traits with membership to broad professional categories and explored its interplay with testing results at cognitive/psychological tests.