Omics in Practice: From Raw Data to Results

The field of omics research is expanding and statistical analysis of new omics datasets provide opportunities to gain more understanding of complex biological systems. While the data becomes more and more abundant and available, it is often neglected that these datasets differ in scale, measurement error and distribution. To avoid biases and misinterpretation due to these differences proper preprocessing of the data is crucial. At the same time characteristics of these datasets, such as highly correlated measurements and high dimensional data, provide challenges for both traditional and new methodology. Finding and interpreting relationships between such datasets can be achieved by latent variable techniques, in which few latent variables model high correlations among measurements and reduce dimensionality of the data. These latent variables typically can be interpreted and used as proxies for association with health traits.

A relatively new omics field is the field of Glycomics. Glycans have already been shown to be associated with both genetic and environmental factors. Integrating glycomics with other omics can help in deciphering their role in complex traits and diseases.

In this project we will start with raw glycan data and apply several data preprocessing tools. At the later stages of the project we will use a flexible latent variable method to unravel the complex correlation structure within glycan data and find relationships with other omics, focusing on metabolic health as an outcome.