Genome Wide Association Studies (GWASs) have identified variants associated with disease that promise to deliver insights into disease aetiology. However, because many GWAS variants lie in non-coding genomic regions, it is difficult to define the genes and pathways underlying a GWAS signal. The possibility of linking GWAS variants to molecular traits, combined with the development of high throughput assays, has motivated the mapping of molecular quantitative trait loci (QTLs), genetic associations with molecular traits such as gene expression (eQTLs) and DNA methylation (mQTLs).

The Finland-United States Investigation of NIDDM (FUSION) tissue biopsy study is motivated by the desire to understand the molecular pathogenesis of Type 2 diabetes (T2D), a complex disease where the vast majority of the ~100 independent GWAS loci occur in non-coding regions. To elucidate the molecular mechanisms underlying these signals, we collected skeletal muscle biopsies, a T2D-relevant tissue, from 318 extensively phenotyped individuals who exhibit a range of glucose tolerance levels. From these biopsies, we generated genotype, gene expression, and DNA methylation information, enabling us to directly measure the effects of T2D on molecular traits, and to link non-coding T2D GWAS loci to candidate molecular targets. In this thesis, I present a catalogue of genetic effects on gene expression and DNA methylation. I use this catalogue firstly, to reveal basic biology of the genetic regulators of skeletal muscle molecular traits, and secondly, to identify molecular traits that are relevant to T2D, glycemic, and other complex traits.

In regards to basic biology, I characterise the broader genomic context of QTLs by calculating the enrichment of QTLs in chromatin states across a diverse panel of cell/tissue types. I also identify key skeletal muscle transcription factors (TFs) and classify them as activators or repressors by aggregating the effects of QTLs predicted to perturb TF binding sites. In addition, I characterise the properties of methylation sites associated with gene expression and use inference models to dissect these methylation-expression relationships, classifying cases where the genetic effect is mediated by methylation, expression, or is independent.

I also integrate molecular trait genetics with complex traits. First, I perform a conditional analysis, mapping GWAS variants for T2D and glycemic traits to molecular traits, prioritising disease relevant skeletal muscle molecular traits. Second, recognising QTLs may also be specific to a disease state or environmental context, I leverage the rich phenotyping of participants to map genotype by environment (GxE) effects on gene expression—eQTLs that exhibit effects specific to an environmental context. Altogether, these analyses form a thorough survey of the genetic regulators of skeletal muscle expression and DNA methylation, and provide an important resource for interpreting complex diseases.