Skip to content

Usage

Command-Line Arguments

Structure Definition

  • --length <LENGTH>: Number of residues (Default: 10)
  • --sequence <SEQUENCE>: Specify sequence (e.g. "ACDEF")
  • --conformation <TYPE>: alpha, beta, ppii, extended, random
  • --structure <REGIONS>: Define mixed structure (e.g. 1-10:alpha,11-20:beta)

Physics & Refinement

  • --minimize: Run OpenMM energy minimization (Implicit Solvent)
  • --optimize: Run Monte Carlo side-chain packing
  • --refine-clashes <N>: Iteratively adjust atoms to reduce clashes

NMR Data Generation

  • --gen-nef: Generate NOE restraints (NEF format)
  • --gen-shifts: Predict chemical shifts (SPARTA-lite)
  • --gen-relax: Generate relaxation data (\(R_1, R_2, NOE\))

Multi-Modal Simulation

  • --mode <MODE>: generate (default), cryo-em, saxs, decoys, dataset, ai
  • --n-decoys <N>: Number of models for ensembles/simulations (Default: 10)
  • --drift <DEG>: Conformational drift for ensembles (Default: 3.0°)
  • --resolution <Å>: Target resolution for Cryo-EM maps (Default: 3.0Å)
  • --mrc-output <FILE>: Filename for simulated Cryo-EM map
  • --q-max <q>: Max scattering vector for SAXS profiles (Default: 0.5 Å⁻¹)
  • --saxs-output <FILE>: Filename for simulated SAXS profile (.dat)

Examples

Basic Peptides

# Alpha helix
synth-pdb --length 20 --conformation alpha --output helix.pdb

# Beta sheet
synth-pdb --length 20 --conformation beta --output sheet.pdb

Multi-Modal Ensembles

# Generate Cryo-EM density map (3.5A res) for a 50-model ensemble
synth-pdb --sequence "MQIFVKTLTGK" --mode cryo-em --n-decoys 50 --resolution 3.5 --mrc-output ubiquitin.mrc

# Generate SAXS profile for a flexible ensemble (high drift)
synth-pdb --length 30 --mode saxs --n-decoys 100 --drift 8.0 --saxs-output ensemble.dat

# Generate and visualize SAXS plots (Kratky/Guinier)
synth-pdb --sequence "LKELEKELE" --mode saxs --visualize --plot-type all

Complex Assemblies (Multichain)

# Generate a heterodimer with Chain A (Alpha) and Chain B (Beta)
synth-pdb --sequence "ALA-GLY-SER:VAL-THR-LEU" --structure "1-3:alpha,4-6:beta" --minimize

High-Throughput Dataset Factory

For large-scale AI training, use the dedicated builder script:

# Build a synchronized dataset of 1,000 multi-modal samples
python3 scripts/build_multimodal_dataset.py --n 1000 --output-dir my_dataset