Diversity Matters: Optimal Collision Energies for Tandem Mass Spectrometric Analysis of a Large Set of N-Glycopeptides

Abstract

Identification and characterization of N-glycopeptides from complex samples are usually based on tandem mass spectrometric measurements. Experimental settings, especially the collision energy selection method, fundamentally influence the obtained fragmentation pattern and hence the confidence of the database search results (“score”). Using standards of naturally occurring glycoproteins, we mapped the Byonic and pGlyco search engine scores of almost 200 individual N-glycopeptides as a function of collision energy settings on a quadrupole time of flight instrument. The resulting unprecedented amount of peptide-level information on such a large and diverse set of N-glycopeptides revealed that the peptide sequence heavily influences the energy for the highest score on top of an expected general linear trend with m/z. Search engine dependence may also be noteworthy. Based on the trends, we designed an experimental method and tested it on HeLa, blood plasma, and monoclonal antibody samples. As compared to the literature, these notably lower collision energies in our workflow led to 10–50% more identified N-glycopeptides, with higher scores. We recommend a simple approach based on a small set of reference N-glycopeptides easily accessible from glycoprotein standards to ease the precise determination of optimal methods on other instruments. Data sets can be accessed via the MassIVE repository (MSV000089657 and MSV000090218).

Can We Boost N-Glycopeptide Identification Confidence? Smart Collision Energy Choice Taking into Account Structure and Search Engine

“…One of the influencing factor in some cases is peptide sequence, which was already implied for Byonic score using sce methods. 38 Here, we found the largest impact of sequence for the Byonic score and pGlyco peptide score. For example, the outlier ENGTISR (derived from AGP, ID: P02763 • A1AG1_HUMAN) has well-separated sets of points with different slopes and intercepts.…”

“…34,37,42,43 Regarding the generic workflows, it has been observed that linear optimal CE vs m/z correlation still has large residuals. 38 Some relevant factors have been identified in the literature but always on a limited set of glycopeptides and/or from the view of fragmentation (not from scoring/confident identification). Furthermore, the findings are somewhat contradictory to each other.…”

“…Experiments were performed with the use of inclusion lists based on DDA measurements taken with sce method optimized for our mass spectrometric platform. 38 Altogether four lists were created: two lists for the mixture of the 3 glycoprotein standards and one list for both HeLa and blood plasma samples.…”

Optimum collision energies for proteomics: The impact of ion mobility separation

“…One of the most widely used fragmentation technique is CID, where the applied CE has profound impact on the information content of the obtained MS/MS spectra. In particular, this phenomenon has been widely documented from practical proteomics aspects for both unmodified peptides 39,40 and peptides with post‐translational modifications 41–47 . It was shown to be important to use instrument specific optimal setting to achieve best proteomics performance in terms of identification scores and number of successful identifications 40,48,49 …”

“…In particular, this phenomenon has been widely documented from practical proteomics aspects for both unmodified peptides 39,40 and peptides with post-translational modifications. [41][42][43][44][45][46][47] It was shown to be important to use instrument specific optimal setting to achieve best proteomics performance in terms of identification scores and number of successful identifications. 40,48,49 The present study focuses on the influence of the application of IMS on the optimum CE at the individual peptide level for a large number of peptides.…”