Sugars/polysaccharides that coat almost all cells and proteins in living organisms
Essential for cell-cell signaling and protein function
Unique to cell and protein type
Image from Glytech Inc.
New biomarkers and therapeutic targets – for example, the glycans expressed on cancer cells are unique and different than the glycans expressed on healthy cells. We can use differential expression of glycans to more effectively diagnose specific diseases or better target treatments to diseased cells while minimizing off-target effects
Discovery of novel glycan-binding proteins and the mechanisms by which they infect and invade microorganisms
Image from Munkley 2019
Why are Glycans so Difficult to Sequence?
Unlike for nucleic acids, glycan synthesis is not template-driven
Diverse and branched structure consisting of multiple chains with different lengths and linkages
Limited structural knowledge from mass spectrometry
Image from glycoMATRIX
Lectins are proteins that have high binding affinity to specific glycan motifs
Glycoprofiles can be deduced from the binding activity of lectins
Framework for Lewis Lab's GlycoSeq model
Set of lectins and known glycan binding sites are used to predict the glycoprofiles on other proteins
Image from cecionilab.com
Predicting Glycoprofiles Using GlycoSeq Pipeline
Machine learning model that takes in lectin binding data as input and outputs glycoprofiles
Trained on data simulated lectin binding data derived from a previous paper (Yang et. al. 2015)
Page Leader: Aditi Gnanasekar