PXDesign — JAX AF2-IG binder design

Binder design with JAX AF2 Initial Guess validation — real ipTM / pLDDT / pAE from the AF2 monomer model run in initial-guess mode against the target. Pilot ~30–60 min.

What it is for

Pick PXDesign when AF2 confidence against a defined target matters and you want real ipTM / pLDDT / pAE on every candidate. For hallucination-driven binder design without AF2 filtering use BindCraft, for antibody and nanobody CDRs use RFantibody, and for target structure generation without binder design use BoltzGen.

PXDesign — Ranomics in-house binder-design pipeline. An RFdiffusion-style backbone generator paired with JAX AF2 in Initial Guess mode for fast, accurate scoring (Bennett et al., Nature Communications 2023). Every candidate carries real ipTM, pLDDT, and pAE from the AF2-IG stage.

When it fits:

  • AF2 confidence against a defined target matters and you want real ipTM / pLDDT / pAE on every candidate.
  • You want faster scoring than full AF2 multimer (Initial Guess shortcut).
  • You want the same scoring pipeline that drives Ranomics' wet-lab campaigns.

Inputs

You will need:

  • Target structure (.pdb / .cif).
  • Chain ID of the target.
  • At least one hotspot residue.

Each run uses a preset that sets the scale and scope:

Pilot — your target, ~45 min
Real PXDesign run against your uploaded target with AF2-IG validation. Up to 24 candidates with real ipTM/pLDDT/pAE scores; results emailed when complete (~30-60 min on A100-80GB).

Parameters you set on the form:

Hotspot residues
Comma-separated target-chain residues defining the epitope. Click in the 3D viewer to toggle.
Binder length
Residue count for the generated binder. ~40 residues is the validated default; PD-L1 published binders are in this range.
Number of designs
How many candidates to score. Higher counts increase cost and runtime linearly.

Typical runtime:

pilot
30–60 min

How to read the results

Ranked candidates with ipTM, pLDDT, pAE, and downloadable PDBs. Target ipTM ≥ 0.70 on 1–2 of 5 designs for a tractable epitope.

Where a tool reports them, the scores mean:

ipTM
Predicted confidence in the binder to target interface. Higher is better. Aim above roughly 0.7 on a tractable target.
pLDDT
Per-residue confidence in the predicted fold. Higher means the model is more sure of that part of the structure.
i_pAE and pAE
Predicted alignment error, at the interface (i_pAE) or across the whole structure (pAE). Lower is better.

Try these examples

One-click sample inputs that load straight into the run form. Edit any field before submitting.

SARS-CoV-2 RBD (6m0j chain E)
Classic de novo binder design benchmark; targets the ACE2 interface on the spike RBD.
PD-L1 ectodomain (4z18 chain A)
Immuno-oncology target. Hotspots on the PD-1 binding interface.

References

Bennett, N. R., Coventry, B., Goreshnik, I., et al. "Improving de novo protein binder design with deep learning." Nature Communications 14, 2625 (2023). Ranomics in-house pipeline; scoring stage uses AF2 Initial Guess.

Open the PXDesign — JAX AF2-IG binder design form All guides