Meta Analyses of Heterogeneous Omics Data

Meta Analyses of Heterogeneous Omics Data (MAHOD) will establish a framework in which data available in online repositories from previously performed Omics studies (i.e. metabolomics for analyses of small molecule metabolites in biofluids, proteomics for similar studies of proteins, etc.) will be analysed simultaneously using highly interpretable machine learning tools. These tools will reconcile differences in experimental design, demographic compositions, and modality (samples and variables that are not directly comparable). The ultimate goal of this research will be enable new insight into disease states and measure the statistical power based on the wealth of information and replicate samples available online. This will information will be exploitable by researchers in academia and industry, and will benefit the European Research Community by making better use of the data currently available for critical projects, and enable new research directions into diseases that are either poorly treatable or understood. The framework will be evaluated by an analysis of existing data on a rare disease (narcolepsy), and a much more common disease (long COVID) to ensure that the technology scales from a relatively low volume of information, to a much more challenging "big data" problem. This proposal will be carried out by an experienced researcher (ER) who completed their PhD in chemometrics, or the application of multivariate statistics and machine learning to chemical data. The ER is in a relatively unique position to bridge the gap between their knowledge of information and computer science, and domain knowledge related to the instrumentation required to perform Omics studies. The ER will be supervised by a computer scientist, who is an expert in the field of multivariate analysis and algorithm design.