Automatically generate 3D structures from sequences and predict binding, 100% within your secure, trusted CDD Vault environment:
Structural biology and structure based drug design (SBDD) tools are in the new CDD Vault AI+ Module. Leverage AlphaFold2 and DiffDock to predict protein structures and model ligand docking via “in silico” protocols directly with your regular, trusted experimental protocols. Use AI+ folding and docking algorithms on your experimental data, registered generative bioisosteres, and deep learning similar molecules from ChEMBL, SureChEMBL, CDD Public, and Enamine in house and Real Diversity (9.6 million structures representing 9.6 billion virtual structures) compounds from the regular AI module. Collaborative teams can view computational and experimental data side-by-side in a single table, to get the best of both methodologies.
AlphaFold2: Predict Protein Structures with Confidence
The groundbreaking AlphaFold2 algorithm from DeepMind is now built into CDD Vault AI+. You can:
- Predict 3D protein structures directly from amino acid sequences.
- Visualize protein models within the Vault interface, alongside associated metadata.
- Export predicted structures for downstream analysis or use them in docking simulations.
Whether you are working with novel targets, exploring orthologs or homologs, AlphaFold integration accelerates structural insights at early discovery stages.
DiffDock: AI-Powered Ligand Docking
CDD Vault now supports DiffDock, a diffusion-based method for fast and flexible small-molecule docking:
- Dock ligands to AlphaFold-predicted or experimental protein structures within the same platform.
- Receive ranked docking poses with confidence scores, enabling quick triage of potential hits.
- Compare predicted poses visually and evaluate key interactions with binding sites.
DiffDock brings deep learning innovation to molecular docking, offering improved accuracy without the need for hand-tuned scoring functions.
Seamless Integration in CDD Vault
Both tools are fully integrated with your existing CDD Vault, and generate docking poses and scores on demand (with zero deep learning expertise required):
- Generate deep learning folding and docking protocols, just the same as your familiar method for generating experimental protocols.
- Predict and dock directly from compound and protein records.
- Store results in your Vault and link them to assays, compounds, and projects.
Keep all your structure-based design data secure, organized, searchable, and collaborative.
Why This Matters
This integration empowers your team to:
- Leverage AI tools without needing specialized expertise, hardware, or external software.
- Enhance structure-based drug discovery with interactive predictions.
- Accelerate project timelines from target identification to lead optimization.
Ready to try it?
The AlphaFold2 and DiffDock features are now available to all CDD Vault users with AI+ (AI plus folding and docking).
For each AI+ package, collaborative teams of 10 researchers will get 750 credits for folding proteins and/or docking small molecules which can be used by any member in your secure CDD Vault. Credits can be shared, so in a group of 10 members, one member could run all 750 foldings/docking and collaboratively share the results with the other nine members in your CDD Vault. The technologies are evolving rapidly, so this policy may evolve over time. Researchers can use as many or few packages of AI+ as they want to acquire.
For more information or a walkthrough or just to try it on your data, contact your CDD Account Manager or email support@collaborativedrug.com.
Other posts you might be interested in
View All Posts
CDD Vault Updates
4 min
April 25, 2025
CDD Vault Update (April 2025 #3): AI+ Folding and Docking
Read More
CDD Vault Updates
10 min
April 18, 2025
CDD Vault Update (April 2025 #2): Pharmacokinetic (PK) and Michaelis-Menten Kinetics (Km/Kd) Curve Fit Equations, Donut Charts, TIFF Image Previews, and Parallel Reactions
Read More
CDD Blog
3 min
April 14, 2025
Let’s Talk Security - Why a Bug Bounty May Be More Valuable Than a Penetration Test
Read More