Protein Design Workflow

Protein Design Workflow Guide

Standard binder design pipeline

Overview

Copy code
Target Preparation --> Backbone Generation --> Sequence Design
         |                     |                     |
         v                     v                     v
    (pdb skill)          (rfdiffusion)         (proteinmpnn)
                               |                     |
                               v                     v
                        Structure Validation --> Filtering
                               |                     |
                               v                     v
                         (alphafold/chai)      (protein-qc)

Phase 1: Target preparation

1.1 Obtain target structure

bashCopy code
# Download from PDB
curl -o target.pdb "https://files.rcsb.org/download/XXXX.pdb"

1.2 Clean and prepare

pythonCopy code
# Extract target chain
# Remove waters, ligands if needed
# Trim to binding region + 10A buffer

1.3 Select hotspots

• Choose 3-6 exposed residues
• Prefer charged/aromatic (K, R, E, D, W, Y, F)
• Check surface accessibility
• Verify residue numbering

Output: target_prepared.pdb, hotspot list

Phase 2: Backbone generation

Option A: RFdiffusion (diverse exploration)

bashCopy code
modal run modal_rfdiffusion.py \
  --pdb target_prepared.pdb \
  --contigs "A1-150/0 70-100" \
  --hotspot "A45,A67,A89" \
  --num-designs 500

Option B: BindCraft (end-to-end)

bashCopy code
modal run modal_bindcraft.py \
  --target-pdb target_prepared.pdb \
  --hotspots "A45,A67,A89" \
  --num-designs 100

Output: 100-500 backbone PDBs

Phase 3: Sequence design

For RFdiffusion backbones

bashCopy code
for backbone in backbones/*.pdb; do
  modal run modal_proteinmpnn.py \
    --pdb-path "$backbone" \
    --num-seq-per-target 8 \
    --sampling-temp 0.1
done

Output: 8 sequences per backbone (800-4000 total)

Phase 4: Structure validation

Predict complexes

bashCopy code
# Prepare FASTA with binder + target
# binder:target format for multimer

modal run modal_colabfold.py \
  --input-faa all_sequences.fasta \
  --out-dir predictions/

Output: AF2 predictions with pLDDT, ipTM, PAE

Phase 5: Filtering and selection

Apply standard thresholds

pythonCopy code
import pandas as pd

# Load metrics
designs = pd.read_csv('all_metrics.csv')

# Filter
filtered = designs[
    (designs['pLDDT'] > 0.85) &
    (designs['ipTM'] > 0.50) &
    (designs['PAE_interface'] < 10) &
    (designs['scRMSD'] < 2.0) &
    (designs['esm2_pll'] > 0.0)
]

# Rank by composite score
filtered['score'] = (
    0.3 * filtered['pLDDT'] +
    0.3 * filtered['ipTM'] +
    0.2 * (1 - filtered['PAE_interface'] / 20) +
    0.2 * filtered['esm2_pll']
)

top_designs = filtered.nlargest(50, 'score')

Output: 50-200 filtered candidates

Resource planning

Compute requirements

Total timeline

• Small campaign (100 designs): 8-12 hours
• Medium campaign (500 designs): 24-48 hours
• Large campaign (1000+ designs): 2-5 days

Quality checkpoints

After backbone generation

• Visual inspection of diverse backbones
• Secondary structure present
• No clashes with target

After sequence design

• ESM2 PLL > 0.0 for most sequences
• No unwanted cysteines (unless intentional)
• Reasonable sequence diversity

After validation

• pLDDT > 0.85
• ipTM > 0.50
• PAE_interface < 10
• Self-consistency RMSD < 2.0 A

Final selection

• Diverse sequences (cluster if needed)
• Manufacturable (no problematic motifs)
• Reasonable molecular weight

Common issues

Advanced workflows

Multi-tool combination

1. RFdiffusion for initial backbones
2. ColabDesign for refinement
3. ProteinMPNN diversification
4. AF2 final validation

Iterative refinement

1. Run initial campaign
2. Analyze failures
3. Adjust hotspots/parameters
4. Repeat with insights