Events Calendar
|
|
|
GlycoShape: From Glycoprotein Structures to AI
Monday 25 August 2025, 02:30pm
Ojas Singh, Maynooth University, Ireland
Location : AB2 5B
About Ojas: Ojas is the developer of GlycoShape, an innovative algorithm that uses AI/ML to restore native glycosylation to protein structures from the PDB or AlphaFold. He will present his work and share his inspiring research journey—from chemist to biophysicist, and finally to an AI/ML expert.
Currently pursuing his PhD in Ireland, Ojas has already co-first authored an article in Nature Methods and co-authored another in Nature Communications. This is a fantastic opportunity to hear from a leading young scientist at the intersection of computational and life sciences.
Abstract: Glycans, complex sugar chains attached to most proteins, are essential for processes such as protein folding, immune recognition, and viral infection. Yet they remain underrepresented in structural biology because their flexibility, diversity, and weak experimental visibility make them hard to capture. To address this, we developed GlycoShape1, a database of realistic glycan ensembles derived from molecular dynamics, and Re-Glyco, an algorithm that restores native glycosylation to protein structures from the PDB or AlphaFold2. Together, these tools make it possible to predict where glycans attach to proteins, explore their likely shapes in solution, and reinterpret experimental data with greater accuracy. Building on this framework, we are now developing an AI copilot for structural glycobiology that orchestrates simulation, fitting, and analysis across these tools, with the goal of accelerating glycoprotein research and moving toward autonomous discovery.
References:
1. Ives, C.M.*, Singh, O.*, D’Andrea, S. et al. Restoring protein glycosylation with GlycoShape. Nat Methods 21, 2117–2127 (2024). https://doi.org/10.1038/s41592-024-02464-7
2. Jumper, J., Evans, R., Pritzel, A. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021). https://doi.org/10.1038/s41586-021-03819-2
Abstract: Glycans, complex sugar chains attached to most proteins, are essential for processes such as protein folding, immune recognition, and viral infection. Yet they remain underrepresented in structural biology because their flexibility, diversity, and weak experimental visibility make them hard to capture. To address this, we developed GlycoShape1, a database of realistic glycan ensembles derived from molecular dynamics, and Re-Glyco, an algorithm that restores native glycosylation to protein structures from the PDB or AlphaFold2. Together, these tools make it possible to predict where glycans attach to proteins, explore their likely shapes in solution, and reinterpret experimental data with greater accuracy. Building on this framework, we are now developing an AI copilot for structural glycobiology that orchestrates simulation, fitting, and analysis across these tools, with the goal of accelerating glycoprotein research and moving toward autonomous discovery.
References:
1. Ives, C.M.*, Singh, O.*, D’Andrea, S. et al. Restoring protein glycosylation with GlycoShape. Nat Methods 21, 2117–2127 (2024). https://doi.org/10.1038/s41592-024-02464-7
2. Jumper, J., Evans, R., Pritzel, A. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021). https://doi.org/10.1038/s41586-021-03819-2
website policy
Connect with us
IISER Mohali, Knowledge city, Sector 81, SAS Nagar, Manauli PO 140306
Telefax : 2240266, 2240124
-
+91 - 172 - 2240266
- +91 - 172 - 2240266