Implementation of a workflow for LC-MS data processing in metabolomics

Metabolomic studies global metabolic changes in biological systems. It is growingly used as a tool for identifying biomarkers that are directly linked to diseases or environmental exposures, thereby aiding in the better understanding of the biochemical mechanisms underlying those diseases, potentially shedding light on new drug targets while also assisting in the understanding of the mode of action of a specific chemical on a cell/tissue/organism. As with all "omics," it provides high-dimensional datasets that must be pre- processed before being statistically analyzed. This, however, necessitates specialized computational and statistical knowledge. The goal of this Master's thesis is to evaluate and improve a workflow for analyzing LC-MS data collected for untargeted metabolomics, making it simpler and easier to use for scientists. In that light, we used Workflow4metabolomics, MetaboAnalyst, and the R packages that power these web platforms to design an appropriate workflow and document each step for a better understanding of the algorithm and statistical methods at work. During this overview of tools provided for LC-MS data analysis, we give insight on some improvements that could truly assist non-statisticians in making better use of the available tools in RStudio: (1) During the pre-processing steps, the parameters can be optimized using the IPO R package, and (2) the MetaboAnalystR package requires some changes to the functions so that the graphs are displayed in RStudio rather than saved directly on the computer allowing also to implement complementary analyses. Finally, we used data from the Metnapar project to demonstrate the application of such a workflow. This project's goal is to conduct metabolomic research on natural anti-parasitic agents. Metabolomics will be used as a tool to investigate the mechanism of action of new natural compounds and compare them to existing drugs.