Why MsXelerator is different:

The MsXelerator software will focus on the detection of all significant chromatographic peaks in your samples. Compared to a manual screening, which often takes hours, MsXelerator’s fast algorithms can do this in seconds. Once all peaks have been detected, the software will reduce the full list by removing:

• 13C isotopes, adducts and fragments (Peaks to Components)
• Align a control sample and remove all peaks present in the control sample
• Identify possible metabolite candidate peaks
• Find isotope signatures in complex samples (e.g. Cl, Cl2)
• Use pre-processing to enhance data quality (baseline correction, smoothing, de-spiking, see Figure 2)
• Overlay UV chromatograms in all plots

Figure 1: Fast Peak Detection, potential metabolic modifications are marked.

Peak Detection, Pre-Processing, Control Checking

The MsXelerator system includes novel processing and visualisation steps to reduce the complexity of data by effectively identifying all LC/MS peaks (Figure 1). Find and review important metabolites in seconds. Use control checking between pre- and post-dose samples to make sure that both expected and unexpected metabolites – such as those resulting from hard to predict oxidative cleavages and rearrangements – are detected.

Control Checking can be done graphically or using the preferred Differential Analysis algorithm. A full analysis only takes a few seconds. Time shifts in the control sample are detected and corrected automatically. Plot the exact mass chromatograms of sample and control in overlay. Peaks from the control can be deleted from the results directly (Figure 3).

Figure 2: Filtering raw data by removing spikes. Top - Raw TIC, bottom – TIC after spike filtering. Small peaks are now much better visible.

Figure 3: Differential Analysis. Detect, mark and remove all peaks from a control sample. Plotted are the mass chromatograms of sample and control at m/z 267.Marked is a peak not found in the control sample.

Isotope Pattern Filtering (IPF)

IPeaks is dedicated to finding peaks/components in your data having specific isotopic patterns. These patterns can be “natural” as will be the case in drugs containing e.g. Cl, Cl2, Br or they may have been introduced by using special stable isotope labeling techniques. The latter ones are often used in the field of (Expression) Proteomics.

Peaks having specific isotope patterns or peaks from labeled compounds are found by IPeaks in a fraction of the time compared to a manual screening of the data set. The search will be performed on the raw data. The IPeaks module contains the following features:

• Five different Isotope Matching algorithms, choose from predefined isotope patterns
  Cl, Cl2, Br, 13C, GSH, or create user specified Isotope Patterns
• Accurate mass and Control Checking

Detection of Chlorine containing Metabolites:

IPF can be used to detect all chlorine containing metabolites in complex samples. In the example it was tested on a urine sample contaminated with a large number of polyethylene glycol peaks. Running the MPeaks algorithm on the full scan LC/MS data set, more than 2500 peaks were detected, of which the majority have no relationship with the parent compound. Running IPF and the Cl2 Isotope pattern, only 43 peaks were detected. Figure 5 shows the results from MPeaks and IPeaks in a so-called dot-plot (time/mz contour map). 

Figure 5: Chlorine Isotope Pattern Filtering. Top - full data set peak picking. Bottom, peak picking results using IPF. All peaks from the matrix were effectively removed. Left – Matrix plot of mass chromatograms containing Chlorine. In overlay are plotted the Mass Defect Filtered (MDF) mass chromatograms. About half the number of peaks cannot be detected using MDF.

Analysing series of samples: Species Comparison

The MS Compare module of MsXelerator offers a large number of tools and views to process series of samples and can be used for Batch / Sample comparison, BioMarker Discovery and Metabonomics. Included are:

1. View series of samples simultaneously.
2. Load MS and UV data from 10-20 samples and click on a any UV peak to automatically identifiy the peak based on its mass spectrum. Align UV with MS data for better comparison.
3. Plot Exctracted Ion Currents and Mass Spectra at any resolution.
4. 3-D views, Heatmaps.
5. Define groups of samples and detect unique differences based on scanning all 2D-surfaces.
6. Statistical Analysis of group differences.
7. Click and Identify: TIC/EIC, BPC/EIC or UV/EIC.
8. Pre-process your data including four different alignment methods.
9. Export results to Excel.
10. Principal Component Analysis and Clustering. Get a good overview of your data.

Data in the example are from a Metabolite Profiling study: 12 samples from 6 species, both male and female were analysed. Samples were analysed on two time points (t=0 and t=3 hours).

MS Compare was used to perform the following tasks:

1. Align all chromatograms using a peak matching algorithm
2. Identify, quantify and tabulate all major metabolite peaks
3. Perform univariate and multivariate species comparison (PCA)

Clicking on any of the displayed peaks in the bottom window will identify the selected sample and plot the extracted ion currents in the top window. Any peak can easily be integrated and added to the result table. The table shown uses a semi-quantitative listing (absent – low – medium –strong). Results can also be displayed using percentage or absolute area counts. If present, the user may also select UV traces to perform the analysis.

Figure 6: MS Compare screen, showing aligned Base Peak Chromatograms of 12 samples in overlay

The result table was created manually in about 10 minutes. Only the larger peaks were selected, identified and quantified. In total 17 metabolites were used in this tutorial (threshold 5%).  In Figure 7, a selection of metabolites is plotted as a function of the sample number. Different relative levels can be plotted and compared easily.

Figure 7: MS Compare – levels of 5 metabolite peaks are plotted for all samples

To analyse the differences and similarities between all profiles based on all 17 peaks, one can make use of multivariate techniques like Principal Component Analysis (PCA) or Clustering. Both techniques will analyse the full table in a multivariate way.

In Figure 8 the scores plot is shown. From this plot it can be concluded that the Metabolite profile for Human Male is closest to the Rabbit Female profile. In a similar way all other samples can be compared and visualized together. On closer inspection, it appears that the pc1 axis is correlated most with the parent compound (plotted in the top window). Alternatively, the samples can also be displayed in a cluster map. The majority of samples appear to have similar profiles, but 4 samples seem to have quite different profiles, which is in agreement with the PCA analysis.