RFantibody Walkthrough

EN / 中文

This page explains how the RFantibody module turns a structure-based antibody design request into ranked, downloadable candidates on hosted GPU compute, and how to reopen and audit saved results.

Open module Workflow

Why this belongs in the platform

Right problem shape

RFantibody is useful when we know a target structure and want new VHH/scFv binders around a chosen epitope.

Clear handoff

The module produces structures, sequences, scores, and manifests that can flow into history, candidates, interface validation, and wet-lab review.

Compute isolated

The web/API contract is separated from GPU execution so the production runner can be Modal or a dedicated CUDA worker.

How CDR boundaries are defined

Residue-level scientific contract for Chothia-derived HLT generation masks versus IMGT sequence annotation, with a real h-NbBCII10 example, exact insertion codes, comparison figures, API fields, and mobile rendering rules.

Open boundary guide

Reopen a saved result

1. Select the run

In RFantibody history, click anywhere on a task row. The selected row is highlighted and its result loads in the workspace above.

2. Verify original inputs

Open Original inputs to verify the target, framework, antigen chains, hotspots, CDRs kept, fixed, or generated, generation scale, sampling temperature, RF2 recycles, deterministic mode, and seed.

3. Review and download

Read the decision banner and candidate evidence before downloading the summary, scores CSV, or manifest. Downloads belong to the selected run.

Operational flow

Upload target PDB and choose antigen hotspots.

target.pdb + hotspots

Choose a VHH/scFv framework, then set each CDR to Keep, Fixed sequence, or Generate.

framework + CDR constraint map

For exact CDRs, the platform first assembles and folds a constrained framework; RFdiffusion-Ab then redesigns only Generate regions.

constrained backbone PDBs

ProteinMPNN samples Generate regions while kept and exact-sequence CDRs remain constrained.

constrained FASTA / sequence table

RF2-Ab predicts complexes and filters by pAE, RMSD, and optional ddG.

ranked CSV/JSON

Reopen any saved run from history, verify its original inputs and evidence, then download the review package.

persistent review workspace

Platform mapping

Epitope Picker

Hotspot residue selection

RFantibody

Backbone and CDR design engine

Runs & Candidates

Source-tagged job history and ranked outputs

Interface

Downstream complex validation

Candidates

Reviewed design library

First-pass gates

RF2 pAE

< 10

Interface uncertainty gate

Self-consistency RMSD

< 2 A

Generated vs predicted complex consistency

Rosetta ddG

< -20

Optional binding-energy sanity check

Hotspots

2-8

Pilot range before broad campaigns

Hosted runner and result retention

Available now

RFantibody requests run on hosted GPU compute. Inputs, status, assessments, ranked candidates, and downloadable artifacts remain available in IASO history after you leave the page.

How to interpret a completed run

Treat the verdict as computational triage, not proof of binding. Confirm the selected run's original inputs, inspect every critical metric and missing field, then move promising candidates into downstream Interface and wet-lab review.