Designing antibodies from scratch used to require years of experimental work. Now, with tools like RFantibody and the Nano Helix interface, anyone can go from a protein target to a designed binder in just a few clicks—no coding required.
This guide walks you through the process of using RFAntibody inside Nano Helix, with direct links to the original structures, preloaded templates, and a step-by-step YouTube video tutorial 👉 Watch here.
🧠 What is RFAntibody?
RFAntibody is a deep learning model from the Baker lab, trained to design antibody-like proteins that bind to specific regions (epitopes) on a target protein. It builds on the RFdiffusion backbone and is optimized for frameworks like nanobodies and scFvs. It's especially powerful for creating therapeutic or diagnostic antibodies without requiring wet-lab optimization upfront.
With Nano Helix, RFAntibody is accessible through a streamlined visual interface that walks you through:
- Selecting your target protein
- Choosing the binding site
- Setting truncation
- Picking a scaffold framework
- Running the full design pipeline
🔬 Step-by-Step: Designing Antibodies with RFAntibody
1. Select the Target Protein
Start in Nano Helix's Protein Models → RFAntibody panel.
You can now select any protein loaded in Nano Helix. This is your antibody target—the protein you want to bind.
2. Define the Binding Site
Once the protein loads:
- You'll see it color-coded (purple, blue, and gray) to guide site selection.
- Select the residues you want the antibody to bind with, and just click "Get current selection".
3. Truncate the Protein
RFAntibody works best with target proteins under 500 residues.
Nano Helix offers:
- Automatic truncation suggestions
- Manual fine-tuning of start and end points
Residues kept in the model are shown in blue, truncated parts in gray, and your hotspot binding site stays visible throughout.
4. Choose Your Antibody Framework
Nano Helix includes preloaded scaffolds:
- Nanobody (from the RFantibody paper)
- scFv (single-chain variable fragment)
You can also upload a custom framework.
Once selected, the 3D view will show:
- The antibody scaffold
- Loop regions highlighted for design
By default, recommended loop lengths are selected. You can adjust them or choose which loops to design.
5. Final Preview & Run the Simulation
You'll now see:
- Target protein in surface mode
- Selected binding site residues clearly displayed
- Antibody positioned for interface design
Click "Submit Simulation" to launch the pipeline:
✅ RFAntibody
→ ✅ ProteinMPNN
→ ✅ RoseTTAFold2
The job may take a few minutes.
🧪 Analyzing the Results
Once the simulation is complete:
- Load the designed structures directly into the viewer.
- Remove the truncated sections for clarity.
- Use the AI structure copilot to align the structures—you'll now see how all five designed antibodies bind precisely to your chosen hotspot.
📽️ Watch the full tutorial on YouTube to follow along step-by-step.
✅ Why Nano Helix + RFantibody Is Ideal for Beginners
- No scripting required: Selection, truncation, and loop editing are all visual.
- Guided design process: Color cues, checklists, and default settings help prevent common mistakes.
- All-in-one pipeline: From epitope selection to structure visualization in a single tool.
Ready to try?
Start designing your first antibody today with Nano Helix.
👉 Launch RFAntibody in Nano Helix
🧠 Frequently Asked Questions
What types of proteins can I design antibodies against?
Any structure available via PDB. Ideal targets are well-folded domains with ≤ 500 residues post-truncation.
Can I use my own antibody framework?
Yes! Nano Helix supports uploading your own nanobody or scFv structure.
How long does a design take?
Typically under 30 minutes, depending mainly on protein size and number of samples.
Do I need coding experience?
Not at all. The entire process is point-and-click.