GC Image Users' Guide

GCxGC-MS Data Processing and Analysis

GCxGC-MS File Input and Output

GC Image supports importing GCxGC-MS data from ANDI files and from Agilent MS files. To import GCxGC-MS data contained in an ANDI file (with .cdf extension) or in a Agilent MS file (with .ms extension), select the Import Image option from the File menu or click the Import Image button from the tool bar of the Image Viewer . GC Image prompts the user to specify the source file name (with .cdf or .ms extension) and the target file name. To maintain the original CDF file, the source and target file names should differ. GC Image also prompts for image metadata such as sampling rate, modulation cycle time, start time, and run time. If the image metadata values are imported from the source file, the value fields will have values and be gray. Grayed values may be overridden by double-clicking in the text box.

GC Image then imports the GCxGC-MS data and displays an image of the total ion count (TIC) at each pixel, as pictured in Figure 1. The TIC is the sum of intensities of all mass/z values for the sample. Most of the image visualization, processing, and analysis tools for GCxGC data are available for GCxGC-MS data. Some of the tools specific to GCxGC-MS are described in this chapter.

Figure 1: An image of the total ion count (TIC) values for GCxGC-MS data.

For GCxGC-MS data, the import dialog allows the user to optionally limit the maximum number of values in each mass spectrum. GCxGC-MS analyses generate very large data sets that may exceed the memory capacity of desktop systems, causing very slow processing. One approach to processing GCxGC-MS data more quickly is to reduce the size of the GCxGC-MS data by retaining only the elements with the largest intensity values in each mass spectrum and using sparse matrices to store the reduced mass spectra.

For example, a GCxGC-MS image with 756x800 GCxGC samples and 215 intensity values in each mass spectrum requires nearly one gigabyte of memory (using four byte integers for the mass/z and intensity values) as shown in Table 1. If only the 32 elements with the largest intensity are stored for each mass spectrum, the memory requirement is reduced drastically to about 0.15 gigabytes as shown in Table 2.

Mass/Z values array	756 x 800 x 215 x 4 = 520,128,000 Bytes	496.03 MB
Intensity values array	756 x 800 x 215 x 4 = 520,128,000 Bytes	496.03 MB
Total memory required to store both the arrays:		992.06 MB

Table 1. Memory required to store the full mass spectrum for each sample in a GCxGC-MS image.

Mass/Z values array	756 x 800 x 32 x 4 =   77,414,400 Bytes	73.82 MB
Intensity values array	756 x 800 x 32 x 4 =   77,414,400 Bytes	73.82 MB
Total memory for both the arrays		147.64 MB

Table 2. Memory required to store the 32 values with the largest intensity for each mass spectrum in a GCxGC-MS image.

Of course, the gain in storage and processing efficiency must be weighed against the loss of data. Figure 2 shows the reduction of a mass spectrum to retain only the values with the largest intensity in a sparse representation.

Figure 2: An example of mass spectrum data reduction with sparse representation.

GC Image imports mass/z values in the given number format, which may be integer or floating point. Some operations on mass spectra (such as NIST MS Library Search) require integer mass/z values. The threshold for rounding up fractional values to the next largest integer can be set via Configure -> Multi-Channel. The threshold can be any non-negative value less than 1.0. The default threshold is 0.5, so that numbers with fractional values less than 0.5 are rounded to down and numbers with fractional values greater than or equal to 0.5 are rounded up.

Mass Spectrum Viewer

Figure 3: The Mass Spectrum Viewer GUI.

The tabular view (on the right side) has two columns: one for mass/z values and one for the corresponding intensity values in the mass spectrum. The intensity values can be expressed as absolute values, as percentages of total intensity (so that all values sum to 100), or as relative percentage values (so that the largest valued entry is 100.00). The choice of value type is set by the radio buttons at the bottom of the window. The entries in the table can be sorted in forward or reverse order by clicking on the label at the head of the column to be used as the sort key. (Consecutive clicks reverse the sort order.) At the top of the table is the total intensity count (TIC) for the mass spectrum.

In the graph (on the left side of the window), the mass/z values are on the horizontal axis and the intensity values are on the vertical axis. The range for the horizontal axis is determined by the minimum and maximum mass/z values in all mass spectra of the image (so, the same absolute scale is used for the horizontal axis in all the mass spectra graphs). The domain for the vertical axis is from zero (or the minimum intensity if it is less than zero) to the largest intensity value in the selected mass spectrum. The vertical axis is expressed as intensity value, percentage of total intensity, or relative percentage, depending on the setting of the radio buttons at the bottom of the window.

M/z labels are placed adjacent to selected large-intensity MS values. As the cursor is moved across the graph, the mouse-over tool-tip indicates the mass/z and intensity value, as in Figure 3. Click-and-drag with the right mouse button defines a zoom window for redisplay of a region of the graph. To return to a full view of the graph, click the Reset button.

Mass Spectrum Selection Modes

• Point Mass Spectrum
• Peak Mass Spectrum
• Blob Mass Spectrum

As illustrated in Figure 4, the Image Viewer palette contains a button to set the cursor mode to mass spectrum and a pull-down mass spectrum mode selector.

Figure 4: Mass spectrum mode selection.

In Point Mass Spectrum mode, when the user clicks the left mouse-button on a point in the TIC image (e.g., as in Figure 1), GC Image retrieves the mass spectrum for the corresponding sample and displays the mass spectrum in a popup window, as in Figure 3 .

In Peak Mass Spectrum mode, when the user clicks in a blob in the TIC image, GC Image displays the mass spectrum of the sample within the blob that has the largest total ion count (TIC), i.e. , the mass spectrum at the blob peak. This mode is accessible only after blob detection has been performed.

In Blob Mass Spectrum mode, when the user clicks in a blob in the TIC image, GC Image displays the sum of all of the mass spectra of all the samples in the selected blob. In computing the sum, the mass/z values are rounded to the nearest integer. This mode is accessible only after blob detection has been performed.

Using the Multi-Channel dialog via Configure -> Configure Settings, the user can specify whether successive selections of mass spectra are presented in multiple windows (in which case a new window is opened for each successive mass spectrum) or in a single window (in which case the previous mass spectrum graph is replaced in the same window by the new mass spectrum). The Windows pull-down menu in the Image Viewer allows switching focus between multiple windows that may be open.

Mass Spectrum Ranges

Figure 5: A mass spectrum in the Mass Spectrum Viewer with two subranges selected.

Selected Ion Chromatograms

To generate a SIC, first select a chromatogram (for a pixel, peak, or blob) displayed in a Mass Spectrum Viewer window. Then, in the graphical display area of the Mass Spectrum Viewer, select one or more mass/z subranges. (A subrange may have a single mass/z value.) Finally, click the Open Ion Image button in the Mass Spectrum Viewer control area.

Using the Multi-Channel dialog via Configure -> Configure Settings, the user can specify whether successive ion chromatogram images are presented in multiple windows (in which case a new window is opened for each successive ion chromatogram) or in a single window (in which case the previous ion chromatogram is replaced in the same window by the new ion chromatogram image). The Windows pull-down menu in the Image Viewer allows switching focus between multiple windows that may be open.

Figure 6: An example ion chromatogram image for the mass/z subrange 90.5-91.5.

The SIC Image Viewer provides a Save button, which allows a user to save a SIC image to a separate file. The button invokes a file browser to specify the location and name of the GCI (.gci) file. An accompanying ANDI NetCDF (.cdf) file also is saved.

Characteristic Ions

The Characteristics Ions attribute is specified as list of ion ranges, each separated by semicolon. Each range is given as single number, in which case the integer range around the value is determined by the ion rounding configuration is used, or as two numbers separated by a comma or hyphen, in which case the specified range is greater than or equal to the specified lower bound and less than the specified upper bound. For example, a Characteristic Ions setting could be ""65.8-66.8;71.8-72.8". The Characteristic Ions Response attribute is the semicolon-separated list of computed responses for each of the indicated ranges.

Mass Spectrum Library Search

GC Image interfaces with the National Institute of Standards and Technology (NIST) MS Search Program and through it can search the NIST/EPA/NIH Mass Spectral Library, the NIST/EPA/NIH Mass Spectral Demonstration Library, or any mass spectral library that is format-compliant. The CDROM installs the MS Search Program (Version 2.0) and the NIST/EPA/NIH Mass Spectral Demonstration Library, which contains 2400 compounds, a small fraction of the full NIST/EPA/NIH Mass Spectral Library. For information about purchasing the full NIST/EPA/NIH Mass Spectral Library, click here .

GC Image supports both identity search and similarity search of the designated library (or libraries). If the unknown compound is likely to be in the library, identity search is the quickest way to find it. Identity search uses a very efficient method of selecting a small set of spectra for subsequent spectrum-by-spectrum comparison. The two types of identity search are Quick Search and Normal Search. Similarity search may be used for spectra that probably are not in the library. The four types of similarity search are: Simple, Hybrid, Neutral Loss, and MS/MS.

• Quick Search is a straightforward comparison of submitted spectra to the library spectra. Quick search is more likely to miss distorted spectra, but is several times faster than normal search described below.
• Normal Search applies penalties on the match factors depending on the rarity of the compound and so it never fails in retrieving matched compounds even when the submitted mass spectrum is distorted. Normal search positions common compounds above exotic isomers that have nearly identical spectra in the result set (hit list).

The NIST Mass Spectral Search Program returns a "hit list" of matched chemical compounds from the library and several factors for each matched compound. GC Image displays these in a table, as illustrated in Figure 7. Compounds in the hit list are listed with the following attributes:

• Compound Name is the primary name in the library. Synonyms are not listed.
• Formula is the molecular formula.
• The Match Factor (MF) is the normalized dot product with square-root scaling of the submitted mass spectrum and a library mass spectrum, using all the elements in the submitted mass spectrum.
• Reverse Match Factor (RMF) is the normalized dot product with square-root scaling of the submitted mass spectrum and the library mass spectrum, but the elements that are not present in the library mass spectrum are not included.
• Probability is the estimated relative likelihood of that the compound mass spectrum is the correct match for the submitted mass spectrum.
• CAS Number is the registry number of the matched compound in the library.
• Molecular Weight is the molecular weight of the matched compound.
• Library is the name of the library in which the matched compound was located.
• Library ID is the identification number of the matched compound in the library.
• NIST# is the NIST identification number of the matched compound.

Figure 7: Example MS Search Hit List Table.

By default, the hit list contains ten compounds with the highest match factor. The user can configure the size of the hit list using the Library Lookup panel in Configure Settings, illustrated in Figure 8. The user also can designate the library to be searched and the type of identity search.

Figure 8: Library search configuration pane.

Library Search for Blob Identification

Figure 9: Blob Metadata popup with a button to access the hit list.

Figure 10: Example MS Search hit list for a blob.

The library search be made on all blobs or on all selected blobs and the search can be made using either the peak mass spectrum or the integrated mass spectrum of each blob (configured by the user, as illustrated in Figure 8 ). The user also can set the configuration so that library search is performed automatically after blob detection.

CLIC expressions can combine MS library search and other MS-related functions with retention-time templates and graphical objects. CLIC expressions and the CLIC tool are described in Chemical Identification.

Contents       Previous: Peak Detection and Analysis       Next: Chemical Identification

GC Image™ Users' Guide © 2001–2007 by GC Image, LLC, and the University of Nebraska.