Using the identified peptides, each LC-MS/MS dataset was aligned against a master FTICR LC-MS dataset using msalign [20] and merged. All identified peptides with a best Mascot ion score of at least 25 were then aligned against each individual FTICR LC-MS dataset, one for each biological replicate and time point. Using these alignments, the peaks corresponding to the identified peptides were integrated over the duration of the chromatographic peak. The data analysis ARRY-438162 clinical trial workflow is illustrated in Figure 5. Only peptide identifications confirmed by

accurate mass measurement were thus used. The peptides were then grouped into proteins, using only peptides attributable to a single protein, and the sum of all peptide intensities used as a measure of protein abundance. The data was normalized against the most abundant protein and the earliest time point. The resulting relative protein intensities were log2-transformed and visualized using the gplots package in R. In the same package we created hexadecimal color codes corresponding to the average values over all expression ratios for each protein. An expression ratio of +2.5 thus corresponded to #00FF00, 0 to #FFFF00 and -2.5 to #FF0000. The color codes were then mapped onto metabolic pathways 4EGI-1 mouse available in the Kyoto Encyclopedia of Genes and Genomes (KEGG) [21]. Figure 5

Data processing workflow. The data obtained from the FTICR-ion trap cluster was processed using the workflow illustrated here. First, the LC-MS/MS datasets from the ion trap were searched against the Escherichia coli protein sequence database using Mascot. Each individual result was aligned to a single master LC-MS

dataset and then merged into one file with aligned retention times. Each separate FTICR LC-MS dataset was aligned against the merged LC-MS/MS data (and hence the master FTICR dataset). Intensities of the identified peptides were then extracted from each FTICR LC-MS dataset by taking the maximum signal in a window of defined m/z and retention time relative to the identified peptide. The resulting list contained the protein name, peptide sequence, maximum Celecoxib observed ion score, and absolute intensities for each peptide. This information from each sample could then easily be collapsed into a single, uniform sample/data matrix with the total absolute intensities for all identified proteins and samples. Acknowledgements The AZD8931 purchase authors wish to thank René van Zeijl, Hans Dalebout, Hannah Scott for technical assistance and Mao Tanabe for kind help with the KEGG pathway “”mapper”". Electronic supplementary material Additional file 1: Peptides identifications. The file represents peptide identifications obtained after Mascot search of all IT LC-MS/MS data and alignment to master FTICR LC-MS dataset. (XLS 726 KB) Additional file 2: Summarized peak intensities. The file provides absolute intensities for a list of all identified proteins in each experiment at each time point. (XLS 476 KB) References 1.