Differential Profiling of Volatile Organic Compound Biomarker Signatures Utilizing a Logical Statistical Filter-Set and Novel Hybrid Evolutionary Classifiers.

摘要

Volatile organic compounds (VOCs) can be monitored to reveal the identity of a unique individual, as well their physiological status. Given the analysis requirements for differential profiling via gas chromatography/mass spectrometry, our group has developed a novel informatics platform, Metabolite Differentiation and Discovery Lab (MeDDL). MeDDL's toolset identifies candidate VOCs to be used for classification. A K-nearest neighbor classifier and genetic algorithm (GA) are used to optimize the classifier and subset of VOCs. The GA uses the area the ROC curve as the optimization measure. Very promising results have been obtained on over a dozen odor recognition problems.