Expression Proteomics / Differential Analysis

In a typical LC/MS based protein expression profiling experiment, multiple samples collected from different patients are analysed in parallel. Each sample is digested into peptides and subjected to multi-dimensional liquid chromatography for separation. Each peptide fraction is then analysed on an LC/MS system. Due to experimental variations a number of processing steps are necessary. These include:

• Peak Picking to detect significant chromatographic peaks
• De-isotoping and charge convolution
• Alignment of chromatograms between samples
• Normalization and pre-processing
• Multivariate Statistical Analysis of differentially expressed peaks

Peak Detection and Quantification

The MsXelerator system includes novel processing and visualisation steps to reduce the complexity of data by effectively identifying differentially expressed LC/MS (chromatographic) peaks. These include:

• Sensitive and fast algorithms to detect significant peaks in the sample. Either use peakheight or 3D volume under peaks for accurate quantitation.
• De-isotope results and determine charge states (Figure 1).
• Apply Differential Analysis between 2 samples using a local optimized alignment algorithm. 
• Select from 4 alignment algorithms to correct for run-to-run retention time variations across runs (Figure 2).
• Select from a number of pre-processing tools: Baseline Correction, Smoothing, De-Spiking.
• Select from 5 different Normalization techniques or import normalization constants e.g. Total Protein Content.

Results from a Differential Analysis can be combined with:

• Peptide/protein ID search results from Mascot.
• View statistical significant peaks for which no Mascot result exists.
• For every Mascot query, view Full Scan MS data or MS/MS spectrum.
• Sort Mascot results on Score, Query Number, precursor m/z value or intensity. Explore results in MsXelerator’s Browser.
• Enhance peptide/protein coverage by analysing differential peptides for which No MS/MS spectrum was measured (directly from raw data). Build and export inclusion or exclusion lists (Figure 3).

Figure 3: Differential Analysis Results comparing two samples. Left: displayed are the exact mass chromatograms of sample (blue) and control (red) for a small differential peak, m/z 883.05  Right: dual display of mass spectra of sample and control. The differential analysis table shows that for this peak no MS/MS spectrum was acquired. In total 603 differential peaks were detected, however, after MS/MS checking it appears that for the majority of peaks no MS/MS spectra were acquired (75%). The problem is caused by the fact that in data dependent MS, the MS/MS is focused on the more abundant peptides and misses the small but interesting peaks!

Figure 4: Mascot Search Results integrated with the with MsXelerator’s Browser. Bottom: Mascot Search list, sorted on Score. Middle: extracted mass chromatogram in either nominal or exact mass mode. Top: MS or MS/MS mass spectrum at selected scannumber.

Statistical Analysis and Visualization

Apply univariate of multivariate statistical methods to detect and validate upregulated or downregulated peaks:

• Detection of differential peaks can be performed using techniques based on peak detection or non-peak-detection algorithms. Non-peak detection methods can be based on the analysis of Mass Spectra or Mass Chromatograms.
• Apply statistical and multivariate methods for validation (t-Tests, ratio’s, PLS-DA, Cross Validation and Bootstrapping).
• Analyse complex Experimental Designs (Time, Dose, Patients) using sophisticated Multivariate techniques.
• Use unsupervised multivariate techniques like Principal Component Analysis (PCA) and Clustering to explore your data before you start.
• View Extracted Ion Currents or Mass Spectra at any resolution.
• Select from a large number of interactive plots to view results: Biomarker Surface map, Profile Plots, Heatmaps, etc.

Figure 5: Overlay of Extracted Ion Chromatograms for a unique feature. Bottom - BioMarker Surface Map showing all detected features in the time – m/z contour map.