Home  /  News & Events  /  Events

Columbia University: Multi-omics Boot Camp: Analysis of Omics Data for Research Studies

Many observational studies are now leveraging modern technologies to measure multiple types of omic data. Omic tools can assist in better characterizing risk factors (e.g. germline genetics, exposomics), providing measurements for intermediate variables that may capture the underlying mechanism (e.g. transcriptomics, proteomics, metabolomics, and the microbiome), or defining an outcome of interest. In this context, the investigator is often confronted with an analytic decision on a continuum between simplicity and complexity. Simple approaches often treat sets of variables in a pairwise independent manner, sacrificing joint evaluation for benefits in interpretability. At the other extreme, complex methods may more-completely model the joint omics structure, but can sacrifice interpretability. This workshop will cover several different approaches to the analysis of multiple omic data, with a focus on the tradeoffs between simple and complex approaches.

This three-day intensive workshop will provide an overview of multiple approaches to analyze multiple omic data types measured on the same individuals or via the use of summary statistic data. Instructors have experience in developing and applying methods for omic analysis in genetic and environmental epidemiology and are members of an active program project focused on developing statistical methods for integrated analysis. The workshop will include seminar lectures with hands-on computer lab sessions to put concepts into practice. Since the analysis of multi-omic data is broad in scope, the workshop will survey a range of approaches and highlight the appropriate application and interpretation of each approach for specific research questions. The lab sessions will provide an opportunity to work hands-on with different types of omic data.

By the end of the workshop, participants will be familiar with the following topics:

Data reduction, including clustering.

Regression analysis, including regularized regression, hierarchical models, and partial least squares.

Interaction analysis and Genome-wide interaction scans (GWIS).

Mediation analysis.

Polygenic/polyexposure analysis.

Integrative analysis using genome-wide association study (GWAS) summary statistics.