Biomedical Engineering Reference
In-Depth Information
The peak list comprises a series of ion chromatograms taken at each mass
channel detected in the chromatogram builder. Some ion chromatograms
may contain more than one peak so a second peak deconvolution stage is
required. Chromatogram deconvolution has several different methods to
choose from, here local minimum search was used. This method searches
for local minima in partially overlapping peaks and works well for
chromatograms with well-defi ned peaks and low noise. Alternative methods
are simple Baseline Cut-Off (threshold), Noise Amplitude (detects baseline
noise amplitude) and Savitsky-Golay (standard peak detection method
using second derivatives). After this step, mzMine produces a resolved peak
list with one peak per row as shown in Figure 4.7. These data can be further
explored by visualisation as a 3D plot (Figure 4.8) or 2D-Gel view. The 3D
view is particularly useful as the detected peaks from a peak list are shown
on the plot, which enables a visual check that peaks have been found
correctly. Alternatively, several peak lists may be obtained by varying the
detection parameters and visualised in the 3D view, the aim being to
recognise the main peaks without excessive detection of baseline noise.
4.5 Metabolomics data processing
In metabolomics a number of samples are measured, resulting in a 3D
LC-MS or GC-MS scan for each sample. Typically, this will consist of
control, treated and repeated standard samples. The aim of metabolomics
processing is to combine these scans together so that the relative amounts
of metabolites occurring in all samples may be compared. The combination
of data has to be done in a consistent way over all the data sets, and
problems have to be accounted for such as small retention time drifts.
4.5.1 Processing a metabolomics data set
in mzMine
Aside from the peak detection and data display features demonstrated
above, mzMine is primarily a tool for metabolomics and has a number of
useful features to support the extensive data processing required. The
batch mode tool allows a chain of processes to be created; a very useful
feature with large data sets as the processing can be set to run overnight
in unattended operation. Furthermore, once the parameters for a
particular operation have been confi gured, mzMine remembers the last
used settings such that they may be applied across all samples in the study.
Search WWH ::
Custom Search