Mass spectrometry technologies offer the promise to comprehensively identify all of the proteins present in complexes, cells and tissues. A typical analysis generates large numbers of peptide (bits of protein) spectra. A search engine then compares the experimental peptide data to theoretical peptide sequences in a protein database and identifies each peptide in the sample. But the search engine often falsely identifies the peptides.
To combat search engine error, Andrew Link, Ph.D., associate professor of Pathology, Microbiology and Immunology, and colleagues have now designed and implemented a novel algorithm called De-Noise. They optimized De-Noise using data sets generated from various control and biological samples and run on different mass spectrometers.
They report in the Journal of Proteome Research that De-Noise improves peptide identification by SEQUEST, one of the most widely used search engines. They demonstrated that De-Noise performs better than other methods used to validate SEQUEST peptide matches. The De-Noise software is available upon request and can be easily implemented with other search engines.
This research was supported by grants from the National Institutes of Health (GM064779, RR024975), by a contract from the National Institute of Allergy and Infectious Diseases and by a Vanderbilt IDEAS Program grant.
