| Resumo: | Metabolomics for the diagnosis of tumor and monitoring treatment is a promising concept since tumor metabolism markedly differs from the metabolism of normal cells. By monitoring changes in metabolite levels, we might be able to detect cancer and better understand cancer pathological processes and progression.Untargeted metabolomics provides an enormous volume of data that cannot be manually analyzed. Thus, before data analysis, NMR and MS spectra must be corrected for all variation caused by experimental variables. This is a critical part in the data mining process as it can drastically interfere with the next step – multivariate statistical analysis (chemometrics). In this review, we will focus on the potentialities of volatile organic compounds (VOCs) to discriminate patients with cancer. For this purpose,we will describe, first, data mining and extracting performed before NMR and MS data analysis. Furthermore, chemometric tools used for pattern recognition and statistical analysis will be described.
|
|---|