TL;DR
A focused sequence-to-structure UI over per-model run endpoints.
Use it / Skip it
Use it when the immediate task is converting a sequence or multimer sequence into a PDB without entering a larger design workflow.
Use when
You have a sequence and need a quick structure for inspection, Interface handoff, or design-around review.
Don't use for
Do not use it for antibody-antigen binding assessment; use Interface for complex prediction.
Inputs
- Backend
- ImmuneBuilder, ESMFold, Boltz-2, or AlphaFold depending on sequence type.
- Sequence
- Single chain or colon-separated multimer where supported.
- Chain context
- VH/VHH chain labels or multimer options depending on backend.
Outputs
- PDB
- Returned or downloadable structure artifact.
- Job metadata
- job_id and status from the model run endpoint.
- Inline preview
- Rendered structure when the returned artifact is available in-page.
Walkthrough
1. Choose backend based on sequence type
Choose backend based on sequence type.
2. Paste a clean amino-acid sequence
Paste a clean amino-acid sequence. Use colon separation only for multimer backends.
3. Run the model endpoint and wait for result
Run the model endpoint and wait for result.
4. Download the PDB or inspect the inline structure
Download the PDB or inspect the inline structure.
Under the hood
- Wraps /api/v1/models/{key}/run.
- Multimer backends can take 10-30 minutes depending on configuration.
Worked example
Pitfalls
- A folded binder alone does not show antigen binding.
- Wrong backend choice can return poor or irrelevant structures.
- Some backends may return files only through download artifacts.