Problems and Solutions in Analysing Large Scale Metabolomics Data

Speaker:

Tim Ebbels

Computational and Systems Medicine, Department of Surgery and Cancer, Imperial College London, London, UK

Title:

Problems and Solutions in Analysing Large Scale Metabolomics Data

Abstract: Large scale studies are becoming increasingly common in metabolomics. However, as sample sizes increase to many thousands, so do the number and complexity of associated challenges. Large studies often combine samples from multiple cohorts of individuals, and may utilise multiple metabolomic assays, both of which lead to problems of data integration. In addition, metabolomic platforms are often untargeted – they do not preselect the metabolites to be assayed – resulting in a significant challenge related to matching features and identifying unknowns. In large scale studies, the potential for analytical artefacts such as batch or drift effects is increased and enhanced quality control procedures are therefore needed to ensure production of reproducible and comparable data. In this talk I will outline some of the analytical and data analysis challenges related to large scale human metabolomics, and how we have addressed these with both analytical and chemometric techniques.