Analysis of 953 Human Proteins from a Mitochondrial HEK293 Fraction by Complexome Profiling
Figure 2
LC-MS/MS data processing and protein profile generation.
Figure 2A provides an overview for the main steps in LC-MS/MS data processing. The acquired mass spectrometry data is used to identify peptides and protein for each gel slice via database searches using the Mascot search engine. Resulting peptide identifications together with the LC-MS data are used as input for the label-free quantitation by the Ideal-Q software that integrates the chromatographic peak surfaces for each peptide from respective extracted ion currents. Quantitative information from all LC-MS/MS analyses is then used to determine the relative abundance for each peptide in every gel slice. The final step in data processing uses the peptide profiles to reconstruct the migration profile for each individual protein in the blue native gel separation. Details for the reconstruction of the protein migration profiles are shown in figure 2B with data for the cytochrome c oxidase subunit 6 C protein. First, all peptide profiles of a protein are used to generate a similarity matrix that contains the Pearson's correlation coefficients between each peptide profile. This information is then used to calculate a similarity score for each peptide which is defined as the sum of all the Pearson's correlation coefficients for the peptide from the similarity matrix. The next step uses a Grubb's outlier test on the calculated similarity scores to discard peptide profiles that poorly correspond with the general peptide migration profile. Finally, the peptide migration profiles are ranked in descending order of their similarity score of which the 5 highest scoring peptide profiles are used to construct the protein migration profile by averaging.