GlycopeptideGraphMS: Improved Glycopeptide Detection and Identification by Exploiting Graph Theoretical Patterns in Mass and Retention Time

From left: Corrine Wan, Dr Terry Nguyen-Khuong and Dr Matthew Choo


Matthew S. Choo1, Corrine Wan1, Ce Huang Poo1, Pauline M. Rudd1,2 and Terry Nguyen- Khuong1

1 Bioprocessing Technology institute, Agency for Science, Technology and Research (A*STAR), Singapore
2 National Institute for Bioprocessing Research and Training, Ireland

Published in Analytical Chemistry 2019 91(11): 7236-7244 (Online Version)


Social networks are the last thing that you would expect peptides to have. However, a recent cross-disciplinary effort by the Analytics Group at BTI spearheaded by Drs. Matthew Choo and Terry Nguyen-Khuong have discovered that one class of peptide actually has a kind of social network when analysed using mass spectrometry (MS). This special class of peptide is the glycopeptide—a peptide attached to chains of oligosaccharides called glycans. The work, published in the prestigious journal Analytical Chemistry, explains that mathematical rules can be used to group glycopeptides into networks of interconnected "friends". These networks are mathematical constructs known as graphs and this branch of mathematics is called Graph Theory, giving rise to the name GlycopeptideGraphMS.

Surprisingly, previously unidentified glycopeptides were automatically grouped with identified glycopeptide “families/networks”. This action of grouping the unknown glycopeptides with the known glycopeptides allowed the researchers to confidently identify those unknown glycopeptides. In fact, thrice the number of glycopeptides (>500) were identified from a cancer glycoprotein called the AXL receptor tyrosine kinase. Tripling the number of identifications creates new angles from which this cancer protein can be targeted by immunotherapy and drugs.

The key advantage of GlycopeptideGraphMS is that confident identifications can be made without heavy reliance on the performance of mass spectrometer’s ability to fragment these biomolecules. Current methods rely exclusively on high quality fragmentation data to validate a glycopeptide. Through the GlycopeptideGraphMS approach, creating networks of glycopeptides greatly improves confidence and even in complex samples. Importantly, this approach also overcomes weakness in the conventional database search approach. The false discovery rate (the number of wrongly identified peptides) was less than 2.21%.

The Analytics group, headed by Dr. Terry Nguyen-Khuong, specializes in analysis of glycans and glycoprotein biologics using mass spectrometry. They also provide a range of services to industry such as evaluating IgG biosimilars, QC screening of biologics and customized LCMS software.