Sigma-Profile Predictor

A message-passing neural network (MPNN) model for predicting molecular sigma-profiles from SMILES strings.

---

Installation

Requirements

Component	Version
Python	3.8 (tested); <=3.11
PyTorch	2.0.x
CUDA	11.7 (recommended); 11.8 or 12.1
PyTorch Geometric	>=2.5.0

Step 1: Create virtual environment

conda create -n sigma-profile python=3.8
conda activate sigma-profile

Step 2: Install PyTorch

CUDA 11.7 (recommended, tested):

pip install torch==2.0.0+cu117 -f https://download.pytorch.org/whl/torch_stable.html

CUDA 11.8:

pip install torch==2.0.0+cu118 -f https://download.pytorch.org/whl/torch_stable.html

CUDA 12.1:

pip install torch==2.0.0+cu121 -f https://download.pytorch.org/whl/torch_stable.html

CPU only:

pip install torch==2.0.0

Step 3: Install PyTorch Geometric

CUDA 11.7:

pip install torch-scatter torch-sparse torch-cluster torch-spline-conv pyg-lib -f https://data.pyg.org/whl/torch-2.0.0+cu117.html

CUDA 11.8:

pip install torch-scatter torch-sparse torch-cluster torch-spline-conv pyg-lib -f https://data.pyg.org/whl/torch-2.0.0+cu118.html

CUDA 12.1:

pip install torch-scatter torch-sparse torch-cluster torch-spline-conv pyg-lib -f https://data.pyg.org/whl/torch-2.0.0+cu121.html

CPU only:

pip install torch-scatter torch-sparse torch-cluster torch-spline-conv pyg-lib -f https://data.pyg.org/whl/torch-2.0.0+cpu.html

Step 4: Install other dependencies

pip install pandas numpy scikit-learn rdkit

Step 5: Verify installation

python -c "import torch; print('PyTorch:', torch.__version__); print('CUDA:', torch.cuda.is_available())"

---

Usage

See USAGE.md for detailed instructions on prediction and training.

---

Data and Models

The dataset and pretrained models are included in this repository:

• Dataset: data/sigma-profile.csv
• Models: models/

---

Project Structure

sigma-profile-predictor/
├── src/                # Library code
│   ├── __init__.py
│   ├── featurizer.py   # Molecular featurization
│   ├── model.py        # MPNN model definition
│   └── predict.py      # Prediction logic
├── data/               # Dataset
│   └── sigma-profile.csv
├── models/             # Pretrained models
│   ├── mse.pt
│   └── composite.pt
├── training.py         # Training script
├── predict.py          # Prediction script
├── requirements.txt    # Python dependencies
├── README.md           # This file (installation)
└── USAGE.md            # This file (usage instructions)