Determining and enhancing metabolite fitness for metabolomics measurements

点击翻译按钮获取中文摘要

Project Summary / Abstract Quantitative measurement of metabolites through high-throughput metabolite profiling technologies is maturing as a clinical and laboratory technique. Reliable results from such metabolite analysis however depend on ensuring that the samples being analyzed are `fit' for analysis. Metabolites are a diverse set of chemical molecules, and are subject to various degrees of susceptibility to chemical, enzymatic and microbial degradation and/or production. Changes in metabolite concentration due to extrinsic factors such as sample preparation, storage, handling etc introduce unwanted noise that limits the power of these multiplexed studies. For example, there is some overlap in metabolites subject to degradation and those reported in cancer biomarker studies, such as glutamate. While several studies have shown that degradation does alter observed metabolite concentrations in human plasma and urine, there are currently no quantitative tools publically available which allow for an assessment of sample fitness. Furthermore, comprehensive information on degradation processes are limited, and do not capture the range of sample types commonly used in oncology metabolomics research. As a result, the foci of the proposed research are to 1) catalogue changes which occur in samples as a function of degradation in various biospecimen types; 2) create a software tool (SAMPLES) which will be made available to the scientific community to assess the extent of sample change independent of biological variation, and 3) examine the extent and type of possible microbial contamination and examine alternative approaches for sample preservation. We propose to examine biofluids particularly relevant to biomarkers studies in oncology research, namely: human serum, human urine, and human plasma (EDTA and heparin anti-coagulants). The outcome of quantitative modeling will be a sample fitness score to assess the integrity of the samples. These tools will be implemented in an open-sourced package for the R statistical computing language framework. We further propose to examine the suitability of several alternate biocides to the conventional approach of adding sodium azide for sample preservation as a means of simultaneously testing the R package and potentially being able to provide enhanced sample fitness to future sample collection. Finally, we propose to make this tool applicable to studies using common analytical platforms including nuclear magnetic resonance spectroscopy and mass spectrometry.