Use of Bayesian Statistics for Pairwise Comparison of Megavariate Data Sets: Extracting Meaningful Differences between GCxGC-MS Chromatograms Using Jensen–Shannon Divergence

详细信息    查看全文

• 作者：Andrei Barcaru ; Gabriel Vivó-Truyols
• 刊名：Analytical Chemistry
• 出版年：2016
• 出版时间：February 16, 2016
• 年：2016
• 卷：88
• 期：4
• 页码：2096-2104
• 全文大小：402K
• ISSN：1520-6882

文摘

A new method for comparison of GCxGC-MS is proposed. The method is aimed at spotting the differences between two GCxGC-MS injections, in order to highlight the differences between two samples, in order to flag differences in composition, or to spot compounds only present in one of the samples. The method is based on application of the Jensen–Shannon divergence (JS) analysis combined with Bayesian hypothesis testing. In order to make the method robust against misalignment in both time dimensions, a moving-window approach is proposed. Using a Bayesian framework, we provide a probabilistic visual map (i.e., log likelihood ratio map) of the significant differences between two data sets consequently excluding the deterministic (i.e., “yes” or “no”) decision. We proved this approach to be a versatile tool in GCxGC-MS data analysis, especially when the differences are embedded inside a complex matrix. We tested the approach to spot contamination of diesel samples.