Fold&Dock

Fold&Dock - a rapid modeling tool for antibody-antigen and nanobody-antigen complexes.

for citations, please cite our paper: End to end accurate and high throughput modeling of antibody antigen complexes

How to run Fold&Dock from google Colaboratory:

1. Open the Colab notebook (Fold_Dock.ipynb, link above).

2. Input antibody sequence

    - Select antibody input type (a fasta file or an antibody sequence). 
      Separate the light and heavy chains with ':' in your input sequence.
      To model multiple antibody sequences for a given antigen in a single run upload a fasta file with multiple antibody sequences.
    - Run this cell to upload your fasta file (if chosen this option).

3. Input antigen structure

    - If you want to preform docking to a given antigen structure in addition to antibody folding, select the option 'do_docking'.
    - If you want to preform docking only for specific chains of the PDB file, specify them in the format 'ABC' for chains A,B,C.
    - Run this cell to upload your antigen pdb file (if chosen the option 'do_docking')

4. Advanced settings

    - Select the number of best scoring complexes to create PDB files for (with an antigen this value can be between 0-len(antigen), 
      without an antigen this value can be either 0 or 1)
    - You have the option to relax the structures and reconstruct the side chains using MODELLER. 
      To do so you need a licese key which  can be obtained from here: https://salilab.org/modeller/
    - You have the option to visualize the best scoring model and select the verbose of the program.

5. Saving options

    - You can select the output directory and whether or not you want to save the results to your drive.

6. Run the other cells without changes.

How to run Fold&Dock locally:

1. Clone the git repository : git clone "https://github.com/dina-lab3D/Fold-Dock"
2. Make sure you have the following libraries installed in your environment:

        - timeit
        - logging
        - argparse
        - pandas
        - subprocess
        - scipy
        - numpy
        - abnumber
        - tensorflow (2.4.0 or higher)
        - Bio (1.8.0 or higher)
        - modeller (optional, only if you want to reconstruct the side chains using modeller, requires license - https://salilab.org/modeller/)

3. Run the following command (with python 3):

        python fold_dock.py <antibody fasta file path>

        options:

                -a <antigen_pdb>: pdb file with the antigen structure for docking.
                -c <antigen_chains>: which antigen chains to consider for docking, for example ABC, (default: All chains in the given antigen_pdb file)
                -o <output directory> : path to a directory to put the generated models in, default is './Results'
                -m : run side chains reconstruction using modeller on the structures, default is False. 
                -t <top_n>: number of models to generate for each antibody sequence (0-len(antigen)), default is 5.
                -v <verbose>: whether or not to print the program progress, default is 1 (print). for a quiet run use -v 0.

Approximate running times (creating PDB files top 5 models):

single antibody sequence (minutes)

100 antibody sequences (hours):