Our lab studies human genetic variation that affects the epigenome and gene regulation, and how these variants contribute to complex traits and disease such as type 1 and 2 diabetes.

The research and work in our lab spans four primary areas:

Mapping the epigenome of disease-relevant tissues and cells

The identity and function of specific tissues and cell types relies on non-coding sequences that regulate gene activity. In our work we map the location and function of enhancers in relevant tissues and cells to understand enhancer function in these cells including transcription factor binding and target gene regulation.

For example in a recent study we mapped chromatin state, chromatin accessibility and 3D chromatin architecture in pancreatic islets through which we defined islet enhancer elements and distal target genes of these enhancers, many of which were highly distal to the genes they regulate. Further integration with T2D genetic fine-mapping data and islet expression QTL data revealed the regulatory programs and target genes of T2D risk variants which uncovered molecular pathways disrupted in disease pathogenesis.

Our current studies include mapping changes in the islet epigenome and enhancer function using these assays under different cellular stimuli, studying the epigenome of other disease relevant tissues such as liver, adipose and brain, and understanding genetic effects on chromatin through allelic imbalance and QTL mapping.

Most recently, our lab has helped pioneer generation of accessible chromatin data from single cells from relevant human tissues such as pancreas, pancreatic islets and hippocampus. As part of these studies we have also developed novel pipelines and tools to analyze single cell accessible chromatin including identification of cell types and subtypes, and mapping genetic effects from single cell data. Analyses of single cell data from both non-diabetic and diabetic individuals will enable understanding of the regulatory programs underlying specific cell types within these tissues and changes in these programs in disease states.

Discovery of novel genetic risk factors and disease mechanisms

We perform population-based genetic association and fine-mapping studies of complex disease in order to map the effects of variants on disease risk. Using this information we identify novel risk loci and causal variants for disease, through which we investigate the mechanisms of how variants at individual risk loci function to derive novel disease biology. In addition, we use genetic association data to more broadly understand disease mechanisms and causal relationships for example through genetic correlations, genomic enrichments and Mendelian Randomization.

Developing statistical approaches for integrating human genetic and epigenomic data

We develop novel methods and tools for integration of genetic and epigenomic data to understand the biological function and disease relevance of genetic variation. Some examples of methodological approaches we are developing include: a semi-supervised learning method for sparse training data to classify variants genome-wide for involvement in complex disease using epigenomic features of known risk variants; tools for mapping allelic imbalance in high-throughput sequence data that model genotyping from sequence data directly, reference bias, meta-analysis of imbalance results across assays, and noisy imbalance estimates; tools for predicting target genes of enhancer activity by integrating data across epigenome assays; tools for mapping cell type specific allelic effects of genetic variants from single cell chromatin.

Creating data science platforms for understanding non-coding disease variation

We developed a cloud-based resource based on open source software developed as part of ENCODE to host and dissemenate human epigenome data relevant to genetic studies of type 2 diabetes. As part of this resource we have collected and re-processed sequence data from thousands of epigenome assays and developed tools to intersect genetic variation with relevant epigenome annotations. We have also developed platforms within our system to enable integration with other resources such as the T2D Knowledge Portal (https://t2d.hugeamp.org/).

Diabetes Epigenome Atlas: https://www.diabetesepigenome.org/