First Published: ASMS (Atlanta) 2nd June 2019.
Introduction
Comprehensive two-dimensional GC-MS is a powerful analytical tool that has evolved from technology used mainly in the R&D laboratory to a robust commercially available solution from several manufacturers. As the amount of data from these instruments can be overwhelming, we have developed AnalyzerPro XD as a vendor neutral 2D data processing solution for all chromatographic-MS data. This software uses chromatographic deconvolution and component reconstruction to isolate the components from each sample.
Here we present a novel technique of non-targeted analysis using 2D GC-MS data to determine the preferred sample preparation procedure for the differenation of whisky samples. Four different commercially available whiskies were subject to three different SPME extraction procedures.
Experimental
Each sample/extraction pair had 3 replicates. The GCxGC TOF-MS conditions are shown in Table 1.
The samples for SPME were prepared as follows:
1. 100 microlitres of neat whisky.
2. 100 microlitres of whisky in 900 microlitres of distilled water.
3. 100 microlitres of whisky in 900 microlitres of distilled water saturated with NaCl.
Figure 1 shows the untargeted PCA analysis of all of the components from each sample preparation method
for each whisky. There were 3 replicates of each sample to give 36 data files. This PCA plot shows that the
largest PCA separation was seen between the NaCl saturated samples and so these samples were invesgated
in isolation from the others to determine the differenation between the whiskies. Figure 2. shows the typical
2D TIC plots from each of the 4 NaCl prepared whiskies.
The visual differences between each sample are quite subtle but the automated chromatographic deconvolution, componentisation, library searching and stastical analysis tools in AnalyzerPro XD allows objective differenation between the samples. Figure 3 shows the PCA analysis for the NaCl prepared samples for each of the whiskies. The various stastical plots available such as a Volcano plot for the fold change between the ‘Bow’ whisky and the ‘GFV’ whisky as well as the Average area plots (shown in Figure 4) can also be selected. This makes the interrogation of the changes between samples much simpler. Combined with the automated NIST target library searching allows putatative identities to be assigned to those differences.
Only a small part of the untargeted workflow has been shown here. Comprehensive GCxGC-MS can produce complicated data sets. The use of chromatographic deconvolution and subsequent 2D componentisation using AnalyzerPro XD allows complex data to be mined. The end-to-end solution providing stastical analysis and display of the data assists the analyst.
Acknowledgement
Dr. Bruce King, U.S. Army Edgewood Chemical Biological Center, U.S. Army Research, Development and Engineering Command.