Modls: Post-translational modification localization scoring with automatic specificity expansion

Probability-based localization scoring of fragment mass-spectrum phosphorylation site identifications has become common practice to confirm search engine modification assignments, and indicate the degree of certainty with which they are defined. Localization of modifications other than phosphorylation is also required but is less commonly supported by current tools. These other modifications, such as hydroxylation, may have broad aminoacid specificity, and can be misassigned when the correct specificity is not considered in an MS database search. In addition, localization software is often specific to a particular MS/MS search engine, and cannot be used to localize modifications identified by multiple search engines. ModLS, a new tool within our freely-available Central Proteomics Facilities Pipeline (CPFP), applies a localization scoring method to arbitrary post-translational modifications (PTMs). As well as localising PTMs based on amino-acid specificities are included in the initial search, ModLS can automatically consider additional specificities from UniMod. This can help avoid 'correct modification, incorrect amino-acid' errors which can occur when data is searched using only a subset of PTM specificities. Localization scoring can be performed on the results from any search engine incorporated within the pipeline, or where the output of individual search engines is combined to give increased coverage. We demonstrate the performance of ModLS using a publicly available phosphorylated peptide dataset, showing that it outperforms the recently characterized Mascot Delta Score approach for CID and MSA data, and is comparable for HCD data. In addition, we show the utility of automatic specificity expansion using hydroxylated and methylated peptide data. ModLS is a user-friendly localization tool for arbitrary modifications. Its inclusion within CPFP allows PTM localization to be performed quickly and easily on large or small result sets, from multiple search engines. Specificity expansion, introduced in ModLS, allows misassignments of modifications due to incomplete consideration of specificities to be identified and Minimized. © 2012 Trudgian DC, et al