ReadMe

Code repository for ourpaper "SITA: Predicting Site-specific Immunogenicity for Therapeutic Antibodies"

SITA takes pdb format files as input and predicts the immunogenicity（SITA score） of each aa residue in the antibody structure.

├─bin
│  ├─feature_extractor
│  │  ├─Information
│  └─model
│      ├─model_files
├─example
   └─PDB_data

Environment setup

1. Change directory to the path of this project

cd {your_path_to_SITA}

2. Run following codes in your terminal

conda create -n SITA python=3.7
conda deactivate # (if base environment is activated)
conda activate SITA
pip install -r requirements.txt

Mind that freesasa might have dependencies on C++ compiler, following the instructions provided by itself'll fix the problem.

USAGE

A command line tool is provided to use SITA for immunogenicity prediction. Usage is as follows:

data preparation

SITA accepts following parameters

usage: SITA.py [-h] -p PDB -c CHAIN -o OUT [-v]

optional arguments:
  -h, --help            show this help message and exit
  -p PDB, --pdb PDB     path to your PDB file folder
  -c CHAIN, --chain CHAIN
                        path to your chain file
  -o OUT, --out OUT     path to the folder you hope to save the output
  -v, --verbose         provides more detailed output if set true

Thus following file/folders should be prepared:

1. pdb file folder: a folder containing all your pdb files. Please check ./example/PDB_data for an example of pdb file folder. Note that structures with only C-alpha are not supported.
2. chain file: a file mentioning the interested chains in the pdb file folder. Data should be orgnized in format {PDB_FILE_NAME},{CHAIN1},{CHAIN2} each line. You might check ./example/PDB_idChain.txt for an example.

running CLI tool

With aforementioned files prepared, run following commands to make predictions with SITA:

cd bin
python SITA.py -p {your_pdb_folder} -c {your_chain_file} -o {your_output_folder} -v