|Home | About | Journals | Submit | Contact Us | Français|
Proteomics data analysis is undergoing a revolution, with new methods of data handling and analysis changing the way that instrumental information is interpreted and validated. This workshop will focus on three main topics: (1) the use of large proteomics data repositories to plan experiments and validate results; (2) the statistical analysis of data from large-scale experiments; and (3) the use of annotated spectrum libraries as an alternative to conventional protein identification software.
Protein identification and characterization by database searching is a well-established technique. It is unlikely that everyone who uses the technique fully understands the statistical nature of the process. Experiments are rarely performed blind, and this can lead to an uncritical acceptance of a result that fits prior expectations. Recently, there have been calls for greater stringency in the reporting of proteomics results. The challenge is for the developers of software tools to improve the technology, so as to reduce the possibility of misinterpretation.
While a spectrum can provide a highly specific fingerprint for a peptide, once identified, that spectrum is often discarded. There has been recent interest in saving these spectra in libraries for rapid and reliable re-identification of these peptides in other analyses. A research program will be described that creates these libraries from highly annotated, quality-controlled “consensus” spectra for reliably identified peptides taken from publicly available MudPIT data-sets. Emphasis is currently on building a comprehensive library of human-derived spectra, which, in collaboration with PeptideAtlas, now include HUPO Plasma and Brain Proteome Projects, as well as other studies of plasma, leukocytes, saliva, and hair. The current version contains spectra for over 45,000 different peptide ions. The potential of this system is described, which is presently limited only by the availability of software and more comprehensive libraries.