BindCraft API
The BindCraft API provides an interface to the BindCraft protein binder design tool. This tool takes as input a target PDB file and generates new novel protein binder PDBs with SolubleMPNN sequences by iteratively predicting and backpropagating through AlphaFold2.
Command Line Interface
Examples
Generate 100 binders for an input PDB with selected hotspots
lev engine submit bindcraft input.pdb \
  --chains A,B \
  --target-hotspots A23,A45,B12 \
  --min-length 5 \
  --max-length 15 \
  --num-models 100
Flags
- --advanced-settings(str) (Optional)- Advanced settings in JSON format
 
- --chains(str) (Required)- Chains to design (comma-separated)
- Example:
        - "A,B"
 
 
- --custom-filters(str) (Optional)- Custom filters for design
 
- --gpu-type(str)- Select the GPU type to use.
- Options:
        - t4(default)
- l4
- a100
 
 
- --max-length(int) (Default:- 15)- Maximum peptide length
 
- --min-length(int) (Default:- 5)- Minimum peptide length
 
- --num-models(int) (Default:- 100)- Number of models to generate
 
- --pdb-file(str) (Required)- The path to a PDB file containing the target you want to design binders for
 - You can specify this flag as the first positional argument or with the - --pdb-fileflag
- --target-hotspots(str) (Optional)- Target hotspot residues
- Example:
        - "A23,A45,B12"
 
 
Python Interface
Examples
Generate 100 binders for an input PDB with selected hotspots
from engine import EngineClient
client = EngineClient()
client.authorize()
job_id = client.submit_bindcraft(
    pdb_path="input.pdb",
    chains="A,B",
    target_hotspots="A23,A45,B12",
    min_length=5,
    max_length=15,
    num_models=100
)
Flags
- advanced_settings(str) (Optional)- Advanced settings in JSON format
 
- chains(str) (Required)- Chains to design (comma-separated)
- Example:
        - "A,B"
 
 
- custom_filters(str) (Optional)- Custom filters for design
 
- gpu_type(str) (Optional)- Select the GPU type to use.
- Options:
        - t4(default)
- l4
- a100
 
 
- max_length(int) (Default:- 15)- Maximum peptide length
 
- min_length(int) (Default:- 5)- Minimum peptide length
 
- num_models(int) (Default:- 100)- Number of models to generate
 
- pdb_path(str) (Required)- The path to a PDB file containing the target you want to design binders for
 
- target_hotspots(str) (Optional)- Target hotspot residues
- Example:
        - "A23,A45,B12"
 
 
Outputs
- bindersdirectory- A directory containing filter-passing binder designs from BindCraft.
 
- rundirectories- N directories containing trajectory-specific data for each run requested with --num-models. Contains the filter values for accepted designs infinal_design_stats.csvand final design filter plots inAccepted/Plots.
 
- N directories containing trajectory-specific data for each run requested with