ImmunoStruct

ImmunoStruct enables multimodal deep learning for immunogenicity prediction

Table of Contents

1. About The Project
   • Key Features
   • Built With
2. Citation
3. Getting Started
   • Prerequisites
   • Installation
4. Usage
5. Model Architecture
6. Troubleshooting
7. Contributing
8. License
9. Contact
10. Acknowledgments

About The Project

ImmunoStruct is a multimodal deep learning framework that integrates sequence, structural, and biochemical information to predict multi-allele class-I peptide-MHC immunogenicity. By leveraging multimodal data from 26,049 peptide-MHCs and jointly modeling sequence and structure, ImmunoStruct significantly improves immunogenicity prediction performance for both infectious disease epitopes and cancer neoepitopes.

(back to top)

Key Features

• Multimodal Integration: Combines peptide-MHC protein sequence, structure, and biochemical properties
• Novel Cancer-Wildtype Contrastive Learning: Enhances specificity for cancer neoepitope detection
• Enhanced Interpretability: Provides insights into the substructural basis of immunogenicity

(back to top)

Citation

If you use ImmunoStruct in your research, please cite our paper:

@article{givechian2024immunostruct,
  title={ImmunoStruct: a multimodal neural network framework for immunogenicity prediction from peptide-MHC sequence, structure, and biochemical properties},
  author={Givechian, Kevin Bijan and Rocha, Joao Felipe and Yang, Edward and Liu, Chen and Greene, Kerrie and Ying, Rex and Caron, Etienne and Iwasaki, Akiko and Krishnaswamy, Smita},
  journal={bioRxiv},
  pages={2024--11},
  year={2024},
  publisher={Cold Spring Harbor Laboratory}
}

(back to top)

Getting Started

To get ImmunoStruct up and running locally, follow these steps.

Pre-requisites

Before installation, ensure you have:

• Python 3.10+
• CUDA-compatible GPU (recommended)
• Conda package manager
• Weights & Biases account for experiment tracking

Dependencies

• python 3.10
• torch 2.1.2
• dgl
• torch_geometric 2.5.3

Installation

1. Clone the repository

   git clone https://github.com/KrishnaswamyLab/ImmunoStruct.git
   cd ImmunoStruct

2. Create and activate conda environment

   conda create --name immuno python=3.10 -c anaconda -c conda-forge
   conda activate immuno

3. Install core dependencies

   conda install cudatoolkit=11.2 wandb pydantic -c conda-forge
   conda install scikit-image pillow matplotlib seaborn tqdm -c anaconda

4. Install PyTorch

   python -m pip install torch==2.1.2 torchvision==0.16.2 torchaudio==2.1.2 --index-url https://download.pytorch.org/whl/cu118

5. Install DGL

   python -m pip install dgl -f https://data.dgl.ai/wheels/torch-2.1/cu118/repo.html
   python -m pip install torchdata==0.7.1

6. Install PyTorch Geometric and related packages

   python -m pip install torch-scatter==2.1.2+pt21cu118 torch-sparse==0.6.18+pt21cu118 torch-cluster==1.6.3+pt21cu118 torch-spline-conv==1.2.2+pt21cu118 torch_geometric==2.5.3 numpy==1.26.3 -f https://data.pyg.org/whl/torch-2.1.2+cu118.html

7. Install additional packages

   python -m pip install graphein[extras]
   python -m pip install lifelines
   python -m pip install -U phate
   python -m pip install multiscale-phate

8. Set up environment variables (if needed)

   export LD_LIBRARY_PATH=/path/to/conda/envs/immuno/lib:$LD_LIBRARY_PATH

(back to top)

Usage

Data Preparation

Place the following files in the data/ folder:

• cedar_data_final_with_mprop1_mprop2_v2.txt
• complete_score_Mprops_1_2_smoothed_sasa_v2.txt
• HLA_27_seqs_csv.csv

Additionally, ensure you have these folders:

• graph_pyg_Cancer
• graph_pyg_IEDB

Generate PyG graph files:

These PyG graph files can be generated using the below command from the corresponding AlphaFold folders.

python immunostruct/preprocessing/cancer_graph_construction_new_KBG.py

Training and Testing

1. Set up Weights & Biases

   Create a project on Weights & Biases matching your project name.

2. Run Experiments

   # HybridModelv2 with full sequence and sequence loss
   python train_PropIEDB_PropCancer_ImmunoCancer.py --full-sequence --sequence-loss --model HybridModelv2 --wandb-username YOUR_WANDB_USERNAME
   
   # HybridModel with full sequence and sequence loss
   python train_PropIEDB_PropCancer_ImmunoCancer.py --full-sequence --sequence-loss --model HybridModel --wandb-username YOUR_WANDB_USERNAME
   
   # Sequence with fingerprint model
   python train_PropIEDB_PropCancer_ImmunoCancer.py --full-sequence --sequence-loss --model SequenceFpModel --wandb-username YOUR_WANDB_USERNAME
   
   # Sequence-only model
   python train_PropIEDB_PropCancer_ImmunoCancer.py --full-sequence --sequence-loss --model SequenceModel --wandb-username YOUR_WANDB_USERNAME
   
   # Structure-only model
   python train_PropIEDB_PropCancer_ImmunoCancer.py --full-sequence --model StructureModel --wandb-username YOUR_WANDB_USERNAME

(back to top)

Troubleshooting

Common Issues

GLIBCXX Error

ImportError: $some_path/libstdc++.so.6: version 'GLIBCXX_3.4.29' not found

Solution: Add your conda environment path to LD_LIBRARY_PATH:

export LD_LIBRARY_PATH=/path/to/conda/envs/immuno/lib:$LD_LIBRARY_PATH

CUDA Compatibility Issues

• Ensure your CUDA version matches the PyTorch installation
• Verify GPU availability with torch.cuda.is_available()

Memory Issues

• Reduce batch size in training scripts
• Use gradient checkpointing for large models

Wandb Authentication

• Login to Wandb: wandb login
• Ensure project names match between script and Wandb dashboard

(back to top)

License

Distributed under the Yale License. See LICENSE.txt for more information.

(back to top)

Contact

Krishnaswamy Lab - @KrishnaswamyLab

Project Link: https://github.com/KrishnaswamyLab/ImmunoStruct

(back to top)