Uploaded October 2026 | Updated November 2026, 2 days ago
Most proteins that cause disease can't be blocked because there's nowhere on them to grab. Molecular glues get around that by deleting the protein instead of blocking it.
Read the paper: nature.com/articles/s41467-026-77657-z
This Tech Talk explains what a molecular glue is, why most disease proteins have no binding pocket, how your cells already destroy unwanted proteins using the ubiquitin-proteasome system, and how a single small molecule can hijack that system by sticking a disease protein to an E3 ligase so the cell throws it away.
We walk through a September 2026 paper in Nature Communications from Baylor College of Medicine, with collaborators including SandboxAQ. The team found a new class of cereblon molecular glues that degrade VAV1, a signalling switch in immune cells, implicated in blood cancers and autoimmune disease. The team also discovered that cereblon grips VAV1 by an unexpected part of the protein, an RT-loop degron rather than the usual G-loop shape. With no 3D structure of the ligase-glue-target complex to work from, the team modelled it using GluePlex, combining AI structure prediction with physics-based calculations, then scored and ranked candidate molecules before any of them were synthesised. The result, NGT-201-18, degraded VAV1 and suppressed activation in human T cells.
CHAPTERS
00:00 The problem
00:44 The cell's disposal system
01:04 How molecular glues work
01:25 Thalidomide, and where glues came from
01:40 The paper: targeting VAV1
02:06 Modeling the complex with GluePlex
02:29 A more potent glue, tested in human T cells
02:40 The takeaway: it's the playbook
THE PAPER
Hanfeng Lin, Xin Yu, Haiyang Zheng, Ran Cheng, Min Zhang, Xiaoli Qi, Yen-Yu Yang, Shengmin Zhou, Rui Qi, Ly Le, Andrea Bortolato, Semen Yesylevskyy, Alan Nafiiev, Xing Che & Jin Wang. "Leveraging high-throughput proteomics and AI-based protein folding to accelerate VAV1 molecular glue discovery." Nature Communications (2026).
nature.com/articles/s41467-026-77657-z
#MolecularGlue #DrugDiscovery #AIforScience
Most proteins that cause disease can't be blocked because there's nowhere on them to grab. Molecular glues get around that by deleting the protein instead of blocking it.
Read the paper: nature.com/articles/s41467-026-77657-z
This Tech Talk explains what a molecular glue is, why most disease proteins have no binding pocket, how your cells already destroy unwanted proteins using the ubiquitin-proteasome system, and how a single small molecule can hijack that system by sticking a disease protein to an E3 ligase so the cell throws it away.
We walk through a September 2026 paper in Nature Communications from Baylor College of Medicine, with collaborators including SandboxAQ. The team found a new class of cereblon molecular glues that degrade VAV1, a signalling switch in immune cells, implicated in blood cancers and autoimmune disease. The team also discovered that cereblon grips VAV1 by an unexpected part of the protein, an RT-loop degron rather than the usual G-loop shape. With no 3D structure of the ligase-glue-target complex to work from, the team modelled it using GluePlex, combining AI structure prediction with physics-based calculations, then scored and ranked candidate molecules before any of them were synthesised. The result, NGT-201-18, degraded VAV1 and suppressed activation in human T cells.
CHAPTERS
00:00 The problem
00:44 The cell's disposal system
01:04 How molecular glues work
01:25 Thalidomide, and where glues came from
01:40 The paper: targeting VAV1
02:06 Modeling the complex with GluePlex
02:29 A more potent glue, tested in human T cells
02:40 The takeaway: it's the playbook
THE PAPER
Hanfeng Lin, Xin Yu, Haiyang Zheng, Ran Cheng, Min Zhang, Xiaoli Qi, Yen-Yu Yang, Shengmin Zhou, Rui Qi, Ly Le, Andrea Bortolato, Semen Yesylevskyy, Alan Nafiiev, Xing Che & Jin Wang. "Leveraging high-throughput proteomics and AI-based protein folding to accelerate VAV1 molecular glue discovery." Nature Communications (2026).
nature.com/articles/s41467-026-77657-z
#MolecularGlue #DrugDiscovery #AIforScience

