Abstract:A critical step in protein biomarker discovery is the ability to contrast proteomes, a process referred generally as quantitative proteomics. While stable-isotope labeling (e.g., ICAT, 18O- or 15N-labeling, or AQUA) remains the core technology used in mass spectrometry-based proteomic quantification, increasing efforts have been directed to the label-free approach that relies on direct comparison of peptide peak areas between LC-MS runs. This latter approach is attractive to investigators for its simplicity as well as cost effectiveness. In the present study, the reproducibility and linearity of using a label-free approach to highly complex proteomes were evaluated. Various amounts of proteins from different proteomes were subjected to repeated LC-MS analyses using an ion trap or Fourier transform mass spectrometer. Highly reproducible data were obtained between replicated runs, as evidenced by nearly ideal Pearson's correlation coefficients (for ion's peak areas or retention time) and average peak area ratios. In general, more than 50% and nearly 90% of the peptide ion ratios deviated less than 10% and 20%, respectively, from the average in duplicate runs. In addition, the multiplicity ratios of the amounts of proteins used correlated nicely with the observed averaged ratios of peak areas calculated from detected peptides. Furthermore, the removal of abundant proteins from the samples led to an improvement in reproducibility and linearity. A computer program has been written to automate the processing of data sets from experiments with groups of multiple samples for statistical analysis. Algorithms for outlier-resistant mean estimation and for adjusting statistical significance threshold in multiplicity of testing were incorporated to minimize the rate of false positives. The program was applied to quantify changes in proteomes of parental and p53-deficient HCT-116 human cells and found to yield reproducible results. Overall, this study demonstrates an alternative approach that allows global quantification of differentially expressed proteins in complex proteomes. The utility of this method to biomarker discovery is likely to synergize with future improvements in the detecting sensitivity of mass spectrometers.

Label-Free Protein Quantification Using Lc-Coupled Ion Trap or Ft Mass Spectrometry: Reproducibility, Linearity, and Application with Complex Proteomes