Periodic Reporting for period 3 - GWAS2FUNC (From GWAS to functional studies: Tackling the complex nature of brain disorders)
Reporting period: 2022-09-01 to 2024-02-29
These disorders have in common a so-called ‘complex etiology': i.e. they are influenced by multiple genetic and environmental risk factors. Each factor contributes only a small proportion to the total disease risk, and each patient potentially carries a different combination of genetic risk factors. The identification of genetic risk factors that are important for brain disorders is therefore complex, and requires studies with huge numbers of individuals included. Additionally, the differences in genetic risk factors between patients for the same disorders complicate the interpretation of the effect of the risk factors and the implications for response to treatment. To advance from genetic factors to treatment implications an important first step would be to understand the biological implications of the genetic risk factors observed for a disorder. Designing functional studies and drawing mechanistic conclusions, which based on the identification of such small-effect risk factors is far from straight-forward.
The primary goal of this ERC project is to bridge the gap between genome-wide association studies (GWAS), which is an approach to identify genetic risk factors, and function. With other words, we aim to improve the understanding of the biological meaning of the genetic risk factors that we are able to link to brain disorders. To achieve this goal an interdisciplinary approach is needed: the fields of human genetics and neuroscience need to work closely in order to design meaningful experiments. This project has three main aims that serve the primary goal:
1. Develop and apply novel algorithms to integrate brain-related biological resources for interpreting GWAS findings.
2. Develop and apply novel algorithms to understand disorder-specific genetic heterogeneity.
3. Assess the viability of iPSC technology in testing functional hypotheses derived from GWAS.
- Review of GWAS in the context of its statistical basis, the analysis procedures, its limitations and challenges, and overview of GWAS applications. This work has been published in Nature Reviews Methods Primers (Uffelmann et al. 2021).
- Review of all resources and methods used by geneticists to point to the most-likely biological mechanism underlying a brain disorder. This work has been published in Biological psychiatry (Uffelmann et al. 2020).
- Dissecting the genetic overlap between anxiety/depression/insomnia by examining overlap in variants, gene(-set) and cell type specificity. Additionally, we aim to study anxiety/depression/insomnia using more biologically grounded phenotypes, that can be more easily translated to animal models. This is accomplished by interaction between genetics and neurobiology scientists. Results on the overlap between anxiety, depression and insomnia are collected in a manuscript that is soon to be submitted.
- Creating generic, easy-to-use scripts for colocalization methods, such that these can be used by the entire department in a consistent manner.
- Study the relationship between depression and atopic disorders (such as allergies, asthma and eczema). There is a moderate genetic correlation between depression and atopy, but Mendelian Randomization did not provide support for a causal relationship. There is moderate genetic correlation between depression and atopic phenotypes, which supports the possibility of a shared biological basis for the disorders.The next step is to a local genetic correlation method (LAVA) to determine which genomic regions contribute to the genetic correlation. These results will then be used to attempt to identify specific biological pathways that might contribute to both depression and atopy.
- Maintaining and upgrading FUMA, an online system for analysis and annotation of GWAS results to help with interpretation and prioritization of GWAS findings.
- We have developed and tested novel algorithms for assessing enrichment for cell-types in disease using a variety of different technologies on the differing genetic architectures. This work is still in progress.
- We have developed a novel algorithm for gene prioritization using information from multiple loci. This has been applied to Insomnia, which has been accepted for publication in Nature Genetics (Watanabe et al., 2022).
- We have applied our current pipeline to brain volume (published in Nature genetics (Jansen et al., 2020)), and Alzheimer's disease (published in Nature Genetics (Wightman et al., 2021) and a second study under review at Communications Biology, and currently available as preprint on medRxiv).
WP2 (Develop and apply novel algorithms to understand disorder-specific genetic heterogeneity)
- Develop an integrated framework for estimating local genetic correlations, which is published in Nature Genetics (Werme et al., 2022).
- Study local genetic sex differences of complex traits, that is using LAVA to estimate local genetic correlations between pairs of complex traits to identify regions that have a correlation unequal to 1 and that can therefore be said to have different effects in males and females. This is done for 150 quantitative traits and the results are of high interest. Currently these results are being interpreted and visualised, and soon the manuscript on this subproject will be written.
- Developing a method to make polygenic risk scores for diseases more interpretable for individuals by transforming them to the probability of disease. We are currently comparing several polygenic risk score methods with regards to how well they are calibrated.
- Using genetic data to examine etiologic heterogeneity in depression. Depression is thought to involve heterogeneity at several levels - symptoms can differ widely between cases, diagnostic criteria can differ between studies, and there may be multiple underlying etiologies that can produce the symptoms commonly recognized as depression. This project focuses on patterns in the co-occurrence of genetic risk factors among individuals with depression, and examine whether these patterns suggest the involvement of different biological pathways in different subsets of depression cases. This project is in a fairly early stage, currently has started to conduct GWAS in which pairs of disorders are grouped together in a single case definition. Some pairs of disorders have clearly distinct etiologies and genetic risk factors, while others are likely to have partially-overlapping etiologies.
- We developed a novel method to assess genetic heterogeneity by assessing Gene by environment interaction using polygenic risk scores. This was applied to Neuroticism and published in Translational Psychiatry (Werme et al,, 2021).
WP3 (Assess the viability of iPSC technology in testing functional hypotheses derived from GWAS)
- Evaluate the iPSC-derived astrocytes. We have previously generated astrocytes from control and schizophrenia iPSC lines. These iPSC-derived astrocytes will be used to study the objectives set out in this project, while also focusing on the mitochondria. To create mature neuronal networks, we are culturing these astrocytes together with patient and control neurons and other glia cells such as oligodendrocytes, obtained from the same iPSC lines. This will allow us to gain mechanistic insight and further evaluate the morphological and synaptic differences by using e.g. immunocytochemistry in an automated high-content setting. We are able to create iPSC-derived astrocytes from SCZ patients which show a normal morphology and become functional when co-cultured with neurons. Currently we are evaluating further the functional effects of co-culturing these astrocytes with neurons with the use of e.g. proteomics and immunocytochemistry.
Novel techniques have been developed and published and are being developed (see above/publications). In the next phase of this project we expect to further develop single-cell RNA methods for GWAS, and possible integration with proteomics data. In addition, we expect to have finalized a new method for gene-prioritization with application to Alzheimer's disease, and psychiatric disorders.
WP2
We expect to have finalized our method for calibration for polygenic risk score for clinical use with application to Schizophrenia and Alzheimer's disease. We currently are also applying machine learning techniques to detect subclusters of patients. This is currently in its infancy, but if successful we will optimize this method for wider use.
WP3
By the end of the project we have a relatively large cohort of iPSC-lines derived from Schizophrenia patients and controls, and if successful we will have more insight in functional and morphological differences between the cells from patients and contols.