Oral Presentation HUPO 2019 - 18th Human Proteome Organization World Congress

Outcomes of the 1st Human Glycoproteomics Initiative (HGI) Study, a Community Evaluation of Informatics Solutions for Intact Glycopeptide Identification (67031)

Morten Thaysen-Andersen 1 , Rebeca Kawahara Sakuma 1 , Krishnatej Nishtala 1 , Kai-Hooi Khoo 2 , Katalin Medzihradszky 3 , Joseph Zaia 4 , Goran Larson 5 , Stuart Haslam 6 , Giuseppe Palmisano 7 , Jong Shin Yoo 8 , Daniel Kolarich 9 , Nicolle H Packer 1 9
  1. Biomolecular Discovery Research Centre, Macquarie University, Sydney, NSW, Australia
  2. Institute of Biological Chemistry, Academia Sinica, Taipei, Taiwan
  3. Mass Spectrometry Facility, The University of California, San Francisco, USA
  4. School of Medicine, Boston University, Boston, MA
  5. Department of Clinical Chemistry and Transfusion Medicine, Gothenburg University, Gothenburg, Sweden
  6. Department of Life Sciences, Imperial College, London, UK
  7. Department of Parasitology, University of Sao Paulo, Sao Paulo, Brazil
  8. Korea Basic Science Institute and Graduate School of Analytical Science and Technology, Chungnam National University, Daejon, Korea
  9. Institute for Glycomics, Griffith University, , Southport, QLD, Australia

Driven by recent analytical advances, LC-MS/MS-based glycoproteomics studies on biological systems now routinely report hundreds and even thousands of unique intact glycopeptides from a single experiment. However, significant bottlenecks still exist in the accurate annotation of the large volumes of MS/MS spectral data and in the correct identification of the corresponding intact glycopeptides. Software for automated glycopeptide identification has become essential. The field has recently seen the development, by both commercial and academic developers, of informatics solutions showing promise for (semi-)automated annotation and identification of intact glycopeptides from MS/MS spectral data. However, their relative performance remains untested.
This first interlaboratory study of the Human Glycoproteomics Initiative (HGI) formed in 2016 under the Human Proteome Project sets out to evaluate the performance of current informatics capabilities for intact glycopeptide identification from complex peptide mixtures. The study comprised 22 participating teams spanning both expert users (13 teams) and software developers (9 teams) across industry and research in the glycoproteomics community. All participants were supplied the same two high-resolution LC-MS/MS datasets of intact N- and O-glycopeptides arising from the analysis of an enriched tryptic digest of human serum glycoproteins. Multiple fragmentation modes (HCD, EThcD, ETciD and CID) were employed. The participants returned detailed reports on the intact N- and O-glycopeptides identified using their chosen informatics strategy. All reports were examined to ensure compliance with the study guidelines and to ensure consistent reporting.
Comparisons of the participant reports provided interesting insights into the diversity of community approaches for glycopeptide analysis, and the group-to-group (vari)ability of accurately identifying intact glycopeptides. The relative performance of the glycoproteomics software and the expert user’s ability to use these tools were evaluated using various tests to assess for correctness of their reported glycopeptides. This presentation will provide an overview of outcomes and conclusions arising from the 1st HGI study.