Can we use DNA methylation to predict disease in diverse populations?

Theme Translational data science

Workstream Omics for prediction and prognosis

Status: This project is ongoing

DNA methylation is a process during which methyl groups become attached to parts of our DNA. Methyl groups are molecules made up of one carbon atom and three hydrogen atoms. When more methyl groups are present near a gene, the activity of that gene is usually turned down or switched off.

Genes are made up of segments of our DNA. While DNA holds the instructions for how our body should be built and maintained, genes carry out those instructions. This means that DNA methylation can regulate when genes work.

Overall, DNA methylation is important because it can regulate how our body functions. However, cancer and other diseases can develop when methylation patterns for certain genes become disrupted.

Using DNA methylation data to predict risk factors and the likelihood of diseases developing is a quickly growing research area. It offers us the possibility of understanding more about our health and providing early interventions for people who may be at risk of developing certain diseases.

However, so far, the DNA methylation data used to train prediction models has come from people with European ancestry. We don’t know yet know whether these models would also work for people with different genetic backgrounds.

Project aims

During this project we will:

  • Test if DNA methylation prediction models trained on populations with European ancestry are generalisable and can be used in people with different genetic ancestry
  • Contribute to addressing issues around the lack of diversity and health inequalities in molecular epidemiology and population health

What we hope to achieve

Improving our understanding of whether prediction tools could be used for diverse groups of people is an important step in reducing health inequalities in the UK and beyond.