When setting up your search in ProteinPilot software, you select either a Rapid Search or a Thorough Search in the Search Effort section. This setting determines which parts of the algorithm will be invoked and effectively how deep into your sample you will search to find the right answers.
When a Rapid Search is selected, on a Fraglet search type will be performed. This type of search is similar to most search engines where the fragment ion spectrum is matched against the possible matches from the database then scored (matches being preselected based on precursor mass).
When a Thorough search is selected, a Fraglet search is performed followed by a Taglet, this type of search is novel and allows for a much more rich definition of the proteome and a much deeper search to be performed. Note when a no-enzyme search is selected, only a Taglet search is performed. Please watch this video on the Paragon Algorithm to learn more details on how this algorithm works
After peptide identification is complete, the Pro Group algorithm assembles the results into a minimal list of detected proteins.
Selection of the search effort also changes the way things like digestion definition and modifications are handled. For example in Rapid mode, the Paragon Algorithm uses simple canonical definitions of proteases, much like other search engines. Trypsin cleaves at Lys and Arg but nowhere else (Left side of the picture). Here, up to 2 missed cleavages allowed.
In Thorough mode, however, the algorithm uses feature probability sets that allow a more refined description of what a given protease does. For the trypsin example, there is a high probability that trypsin will cleave at Lys and Arg, but also we know that we will see semi-tryptic cleavages at other residues at much lower frequencies, and these are described using empirically determined probabilities. With Thorough mode, up to 5 missed cleavages allowed as well as semi-tryptics as shown on the Right of the figure).
The same approach is used for modifications; a much more sophisticated and biologically relevant view of the sample is used when the Thorough search effort is selected. To see which modifications are searched when doing a Rapid search, refer to the Standard workup mod set in column V Unified Modification Catalog.xlsx found in the Help folder. In ProteinPilot software 5.0 version, 63 modifications are searched when Rapid is selected. For a typical search where Thorough is selected with Biological modifications, a much bigger set of modifications is invoked; see the Standard biological mod set in column W (additional 337 mods). By using the Taglet search in combination with feature probabilities, this much-expanded set of mods can be searched with only a small impact on search time and minimal impact on false discovery rates. The Feature probabilities used are the numbers shown in these columns in the Mod Catalogue.
More details on the Paragon Algorithm can be found in this publication. Mol. Cell. Proteomics, 6: 1638-1655, (2007).