|Home | About | Journals | Submit | Contact Us | Français|
Metabolomic data come in many forms depending on what was collected. In a study looking for biomarkers of exposure to H2S, whole blood was collected from participants and subsequently analyzed by GC/ITMS. The data from chromatograms were sorted using principal component analysis. A key portion of the chromatograph was used to demonstrate differences between no exposure, a low and high exposure. True difference mass spectra were used to differentiate urine samples from mice before and after they ocnsumed green tea. MS data from control subjects looking at MTBE exposure sensitivity, showed a time course difference from a single subject, following exposure. All of this data came from mass spectral data points but the pattern may not be appropriate for storage in a database that should be searchable. The patterns created are the most effective way to verify biological differences in conditions and identify potential biomarkers, but is not useful for comparing results from one study to another. Data in the raw phase (mass spectral data) cannot be compared to other raw data (NMR) and cannot be linked directly to metabolic pathways.
A comprehensive database would include data that was; searchable, linked to metabolic pathways and comparable across analytical methods. The ideal database would also contain all requisite study subject data such as; diet, age, gender and genotype. Until we can create the ideal database one relating raw data to either pathway or compound will have to suffice. This presentation will spotlight the different types of data from the studies described above as well as direct infusion MS data. The raw data from mass spectra can be related to compounds in metabolic pathways by generating theoretical mass spectrum. The database storage decisions on what to store and how to store it will be the focus of this presentation.