Dissecting the molecular aetiology of complex traits using high dimensional omic data

The discovery of genetic variants associated with complex traits is increasing at an exponential rate. It is now of vital importance to develop our understanding of the molecular mechanisms which can help explain these findings, in order to improve our capability to prevent and treat disease. Advancements in high-throughput sequencing technologies present an unprecedented opportunity to address this challenge and ascertain the biological and clinical relevance of results from genome-wide association studies (GWAS). However, there is an increasing abundance of data being generated on diverse types of molecular "omic" traits, accompanied by the rapid development of partially overlapping and often untested methodologies. To overcome this challenge, there needs to be focused research into the most appropriate and efficient manner to harness large-scale 'omic data to elucidate the molecular determinants of complex disease.

The research outlined in this fellowship proposal can be delineated into five categories, with the overall aim of harnessing large-scale data to improve patient healthcare in-line with the UK's industrial strategy. The opportunity presented by HDR-UK will allow me to address some of the most crucial limitations in molecular aetiology. Specifically, there needs to be extensive research into tissue-specificity for 'omic traits, systematic frameworks to appropriately appraise molecular mediation and methods to improve causal inference in this paradigm. I also intend on applying novel, state-of-art-methods to available 'omic data to elucidate findings which have translational value for therapeutic evaluation. Finally, I will build publicly accessible computational tools to automate fundamental analyses in this paradigm and resources to disseminate the findings of this fellowship.

Using health informatics to harness large-scale, high throughput data to develop our understanding of the causal pathway from genetic variation to complex disease is the overarching theme of this research project. This research will lead to several high impact publications to improve our understanding of the molecular determinants of disease, as well as web tools that should help disseminate the product of this project and also assist colleagues with their endeavors in this field. As such, this work most closely aligns with the HDR-UK priorities concerning health informatics and accelerating medicines discovery.

1. Tissue-specific Mendelian randomization - As an initial analysis, I will instrument genes using tissue-specific expression quantitative trait loci (eQTL) and assess putative associations using Mendelian randomization (MR) across 139 complex traits. However, as more tissue-centric expression data becomes available I will be able to assess this crucial aspect of molecular aetiology with greater confidence by instrumenting genes with multiple instruments. I also plan to expand upon the initial set of traits assessed using ~600 outcomes from the UK Biobank study as well as proteomic data.

2. Systematic two-step epigenetic MR - Building upon recent research I have led, I intend to construct a framework to assess whether genetic and environmental exposures influence complex traits via changes in DNA methylation levels. To do this I will systematically apply two-step epigenetic MR.

3. Multivariable 'omic MR - I will apply the principals of multivariable MR to assess the effect of multiple 'omic traits simultaneously on complex traits/disease. This work will also evaluate pharmaceutical targets which should be translatable for therapeutic purposes.

4. Application of novel methods to 'omic data - Future methodology in this field is likely to focus on integrative approaches, and therefore I intend to apply the most innovative and cutting-edge methods to multiple types of 'omic data, as well as iterate upon them.

5. Computational tools to disseminate the product of this fellowship - I will develop a suite of bioinformatics tools and resources to disseminate the findings and methods of this fellowship to a broad audience. Building upon experience in developing a target validation tool for industry partners, I plan to develop platforms which automate fundamental analyses in this field to large-scale data which should be of considerable value to the community and industry partners.