Skip to main content
All solutions

Solutions · 5 workflows

Biologics design

Generate binders for your target with diffusion and sequence design, fold-check every candidate, and gate them on quality and developability before you order a gene.

Not sure which? Describe your problem
[Biologics design]Fig. 05
Workflows
5
Stages
5
Tools
10
Each line is one workflow. Lines share a station where they run the same step; select a workflow below the map to jump to its details.
[Biologics design · what goes in, what comes out]

  1. Your target

  2. A nanobody against it

  3. Where it binds

Target
vWF A1 domain
Structure
PDB 7EOW · 1.6 Å
Epitope
16 residues
Shown: the von Willebrand factor A1 domain with caplacizumab, the first nanobody-based medicine, bound to it, from its crystal structure: the 16 residues it touches on the target, the nanobody's 3 CDR loops (the thick lines), and the polar contacts between them (residues are numbered as in the PDB entry). In this family you name the residues to target; the workflow designs candidate binders against them (RFAntibody builds a nanobody's CDR loops), checks that each design folds, and gates the survivors on structure quality and developability. The complex is a real one, not a design from this workflow. Structure: Lee et al. (2021), Biochem Biophys Res Commun.

Workflows

Pick the one that fits.

Open any of them in the builder: signed out you get a read-only preview, signed in you can add your inputs, see the estimate for your plan, and run. The cost cap is the default; you can change it at launch, and a run pauses instead of spending past it.

  • simple5 steps

    Design binders for a target and keep the structurally sound ones

    FreeBindCraft → structure QC → filter

    Prepare your target, generate binders with FreeBindCraft, keep the ten designs the model is most confident in, and drop any whose backbone or side-chain geometry fails the structure quality check. Results are predictions to test, not validated binders.

    You provide
    Protein structure (PDB)
    Runs on
    CPU · A10 GPU
    Cost cap
    $25.00 default
    Open in the builder: Design binders for a target and keep the structurally sound ones (opens in a new tab)
  • simple3 steps

    Rank antibody point mutations by predicted binding energy

    Point mutants → binding ΔΔG → rank

    Make point mutants of your antibody's sequence and rank them by the predicted change in binding energy (ddG) in your antibody-antigen complex. The ranking compares variants with one another within a campaign; it is not an absolute affinity. Up to 50 variants are scored.

    You provide
    Protein structure (PDB), Sequence
    Runs on
    CPU · T4 GPU
    Cost cap
    $15.00 default
    Open in the builder: Rank antibody point mutations by predicted binding energy (opens in a new tab)
  • moderate7 steps

    Design antibodies against a target and screen them for developability

    RFAntibody → QC → developability → humanness

    Design nanobody (VHH) or scFv binders against residues you choose on your target with RFAntibody, drop designs with poor geometry, screen the survivors for charge, hydrophobic patches, and sequence liabilities, then number their chains and keep the five closest to a human germline.

    You provide
    Protein structure (PDB)
    Runs on
    CPU · A100-40GB GPU
    Cost cap
    $25.00 default
    Open in the builder: Design antibodies against a target and screen them for developability (opens in a new tab)
  • moderate6 steps

    Design binders and shortlist a structurally diverse set

    FreeBindCraft → QC → developability → cluster

    Generate binders with FreeBindCraft, gate them on geometry and developability, then cluster the survivors by structure so the designs you order are distinct from one another instead of near-duplicates.

    You provide
    Protein structure (PDB)
    Runs on
    CPU · A10 GPU
    Cost cap
    $25.00 default
    Open in the builder: Design binders and shortlist a structurally diverse set (opens in a new tab)
  • advanced10 steps

    Design nanobodies and check that each sequence folds

    RFAntibody → QC → developability → ESMFold2

    Design nanobody (VHH) binders with RFAntibody and run the geometry and developability gates, then predict the structure of each surviving nanobody from its sequence alone with ESMFold2 and keep the three it folds with the highest confidence (pLDDT 70 or above). Folding the nanobody by itself checks its sequence; it does not predict that it binds.

    You provide
    Protein structure (PDB)
    Runs on
    CPU · A10 GPU · A100-40GB GPU
    Cost cap
    $40.00 default
    Open in the builder: Design nanobodies and check that each sequence folds (opens in a new tab)

Related tools

Run any step on its own.

The tools behind these workflows also run individually.

  • Protein Preparer

    PDBFixer cleanup, PDB2PQR protonation, and dock-ready structures without MD solvation

  • FreeBindCraft

    Design miniprotein and peptide binders against a target structure using open-source scoring

  • Structure QC Report

    Grade protein model geometry: Ramachandran and rotamer outliers, atomic clashes, cis-peptides, C-beta deviation and ligand bond geometry

  • Sequence Mutator

    Apply point mutations such as T315I to a protein sequence, checking each wild-type residue

  • Binding ΔΔG

    Rank antibody variants by predicted change in binding free energy on an antibody-antigen complex

  • RFAntibody

    Antibody and TCR design with RFdiffusion, ProteinMPNN, and RF2 (stages or full pipeline)

  • Protein / Antibody Developability Screen

    Screen protein and antibody chains for pI, charge, GRAVY, instability, and sequence liability motifs

  • Antibody Numbering & Humanization

    Number antibody, nanobody and TCR variable domains (IMGT, Kabat, Chothia), mark CDRs and assign the nearest human germline

  • Protein Structure Ensemble Comparator

    Compare predicted or designed protein structures by TM-score and RMSD, cluster them, and pick one representative per fold

  • ESMFold2

    Predict protein, DNA, RNA, and ligand complex structures with evolutionary scale modeling