Increased sensitivity with automated validation of XL-MS cleavable peptide crosslinks.

Abstract

Peptides crosslinked with cleavable chemical crosslinkers are identified with mass spectrometry by independent database search of spectra associated with the two linked peptides. A major challenge is to combine together the evidence of the two peptides into an overall assessment of the two-peptide crosslink. Here, we describe software that models crosslink specific information to automatically validate XL-MS cleavable peptide crosslinks. Using a dataset of crosslinked protein mixtures, we demonstrate that it computes accurate and highly discriminating probabilities, enabling as many as 75% more identifications than was previously possible using only search scores and a predictable false discovery rate. XLinkProphet software is freely available on the web at http://brucelab.gs.washington.edu. Supplementary data are available at Bioinformatics online.