unSeminar: “Development of methods for multi-omics data integration”

Presented by Danielle Lemay, UC Davis and USDA with colleagues David Mills, Carlito Lebrilla, Carolyn Slupsky, David Dallas & Daniela Barile.

Key Words: data integration, multi-omics, machine learning, network theory, probabilistic graphical models, statistics

Abstract:

There are many ways to potentially integrate multi-omics data with other heterogeneous biological data. This poses a challenge as it is unclear how the different methodologies compare with one another and potentially influence the resulting analyses and inferences from multi-omics studies. This research project aims to provide a benchmark example, either for direct use or to inform simulated data, for assessing different multi-omics data integration methods. Thus far the research team has assembled a combined network from swine milk protein data, which reveals important features and their relationships. However, the evaluation of additional methods and/or the development of adaptable software is needed. The current goal is to develop a pipeline for identification and evaluation of additional methods for the integration and analysis of heterogeneous, multi-omics data sets.